Hidden massive eclipsing binaries in red supergiant systems: The hierarchical triple system KQ Puppis and other candidates

The majority of massive stars are part of binary systems that may interact during their evolution. However, not many RSGs are known binaries, and only a few have constrained orbital parameters. We search the available TESS photometry for eclipsing companions of RSGs. We focus on the best candidate, VV Cephei type binary KQ Pup, which is made up of a RSG, KQ Pup A, and a B-type companion, KQ Pup B (orbital period of 26 yr). We use photometry, spectroscopy, and newly taken interferometric data with VLTI-GRAVITY. Using TESS, we discovered eclipses with a period of $17.2596 : \rm d$, associated with KQ Pup B, making it a Ba+Bb binary. The detection of the hydrogen Br$γ$ line with VLTI-GRAVITY enabled us to track the orbital motion of the Ba+Bb pair relative to A and determine the astrometric orbit of A+B. The dynamical masses agree with independent estimates from asteroseismology and evolutionary models. The results give a mass of $ \sim 10 : \rm M_{\odot} $ for the RSG KQ Pup A and $ \sim 14 : \rm M_{\odot} $ for the sum of the hot components Ba+Bb. We determined an orbital parallax of $π= 1.24^{+0.05}_{-0.04}, \rm mas $, which is the first such parallax measurement for a RSG. KQ Pup represents a unique demonstration of mass transfer mechanism in wide eccentric RSG systems. The variability of Balmer emission lines and the detection of Br$γ$ are a strong signature of accretion to Ba+Bb near periastron. With the RSG filling its Roche lobe only by $\sim 70%$ at periastron, the mass transfer is instead driven by accretion from its extended atmosphere via the Wind Roche Lobe Overflow. The accretion disk dissipates by apastron. Overall, we discovered that several previously assumed RSG binaries host eclipsing inner systems, corresponding to $\sim 10 %$ of all known Galactic RSG binaries. This suggests that many of the remaining RSG binaries may also be hierarchical triples.

💡 Research Summary

This paper presents a comprehensive multi‑technique investigation of hidden inner eclipsing binaries within red supergiant (RSG) systems, focusing on the hierarchical triple KQ Puppis. The authors begin by cross‑matching the catalog of 79 Galactic RSG binaries (Healy et al. 2024) with the TESS full‑frame image database, extracting light curves for 651 RSG candidates. After eliminating blends and nearby eclipsing binaries, six promising systems remain, of which KQ Pup is the most data‑rich.

KQ Pup is a classic VV Cephei‑type wide binary consisting of an RSG (component A) and a B‑type star (component B) on a 26‑year orbit. Using TESS short‑cadence (120 s) data from four sectors, the authors detect a low‑amplitude (≈0.5 %) periodic dimming with a period of 17.2596 days. Detailed period searches (Lomb‑Scargle, Box‑Least‑Squares) confirm that these eclipses cannot involve the RSG itself; instead they arise from an inner binary within the B‑type companion, which the authors label Ba and Bb.

To characterize the system, the team assembled an extensive spectroscopic suite: 66 high‑resolution (R ≈ 55 000) optical spectra from the STELLA echelle spectrograph (April 2024–February 2026), additional high‑resolution spectra from PLATOSpec (R ≈ 70 000), archival IUE ultraviolet spectra, and FEROS/ESPaDOnS data. The spectra reveal variable Balmer emission and He I lines that strengthen near periastron, indicating active mass exchange. Crucially, VLTI‑GRAVITY K‑band interferometry (R ≈ 4000) detects the Br γ (2.166 µm) emission line and provides differential phase measurements for six baselines. By tracking the Br γ centroid across multiple epochs, the authors directly map the orbital motion of the Ba+Bb pair relative to the RSG.

A joint orbital solution is obtained using a Markov Chain Monte Carlo framework that simultaneously fits the astrometric orbit of the outer pair (A + B) and the spectroscopic orbit of the inner pair (Ba+Bb). The outer orbit has a semi‑major axis of ≈13 AU, eccentricity e ≈ 0.42, and an orbital parallax π = 1.24 (+0.05/‑0.04) mas – the first orbital parallax measured for a red supergiant, in agreement with Gaia DR3 within uncertainties. Dynamical masses derived from Kepler’s law are M_A ≈ 10 M⊙ for the RSG and M_Ba+Bb ≈ 14 M⊙ for the hot inner binary, with a lower limit on the secondary component Bb of ≳ 1.2 M⊙.

The authors interpret the mass‑transfer mechanism in the context of the RSG’s Roche‑lobe filling factor, which is only ~70 % at periastron. Direct Roche‑lobe overflow is therefore unlikely; instead, the system exhibits Wind Roche‑Lobe Overflow (WRLOF), where the extended, slow wind of the RSG overfills its Roche lobe and streams toward the Ba+Bb pair. This scenario is supported by the observed enhancement of Balmer and Br γ emission near periastron (phase ϕ ≈ 0.95) and the subsequent weakening around apastron (ϕ ≈ 0.5), indicating the formation of a transient accretion disk around Ba+Bb that dissipates as the stars recede.

Beyond KQ Pup, the survey identifies eclipses in five additional RSG systems, suggesting that roughly 10 % of known Galactic RSG binaries host inner eclipsing companions, i.e., are hierarchical triples. This fraction is comparable to the incidence of triple systems among massive O‑type stars, implying that many RSG binaries may have been mis‑classified as simple binaries due to the low photometric contrast of the inner pair.

In summary, the paper delivers four major contributions: (1) the first orbital parallax for a red supergiant, providing an independent distance scale; (2) the discovery and dynamical characterization of a massive inner eclipsing binary within a B‑type companion, yielding precise masses for all three components; (3) a detailed observational validation of WRLOF as the dominant mass‑transfer channel in wide, eccentric RSG systems; and (4) a statistical indication that hierarchical triples are common among RSG binaries, reshaping our understanding of massive‑star evolution, mass‑loss processes, and the progenitor pathways to core‑collapse supernovae. The work underscores the power of combining space‑based photometry, high‑resolution spectroscopy, and long‑baseline interferometry to unravel the complex architecture of evolved massive stars.