Disco in the Dust: Reflected light at the bow shock around Betelgeuse's companion explains its observed luminosity

Abstract

Recent detections of $α$ Orionis B, the putative companion to Betelgeuse, have been reported using multiple instruments and techniques. These include, most recently, a $>6σ$ detection using VLT/SPHERE reported by Montarges+2026. The authors infer a bright companion ($\sim10^{-3}\times$ Betelgeuse's luminosity) with a mass of $2-3\,M_\odot$ and $T_\text{eff}\approx10-12{,}000$\,K, loosely consistent with the uppermost bounds of Howell+2025's recent mass estimate from speckle-imaging. However, it is in tension by a factor of two with the reported mass-exclusion limits via non-detection from a recent HST far-UV campaign (Goldberg+2025). While mass identification from isochrone-fitting (conducted in all cases) is prone to uncertainty, the restrictive HST upper limit on FUV flux precludes the hot, blue emission of a $\gtrsim2\,M_\odot$ main-sequence star. This discrepancy is reconciled by the following hypothesis: these detections observe the reflection of Betelgeuse's own luminosity scattered off the companion's bow shock and wake as it traverses the dusty circumstellar medium. To evaluate the feasibility of this scenario, we draw from custom, filter-specific stellar models computed with MESA, as well as Athena++ simulations of the companion's bow shock in idealized conditions. Importantly, the expansive bow shock from a stellar-mass object necessarily subtends a sizable portion of Betelgeuse's outgoing flux. We find that optical luminosity ratios of $\sim10^{-3}-10^{-4}$ are straightforward to resolve within reasonable assumptions about the companion mass and circumstellar dust, and disfavor a hot $3\,M_\odot$ companion. We thereby reconcile competing mass hypotheses across observational campaigns through proper attribution of the observed brightness, motivating future multi-wavelength campaigns to further characterize this enigmatic system.