Close-in faint companions mimicking interferometric hot exozodiacal dust observations

Context . Interferometric observations of various nearby main sequence stars display an unexpected infrared excess, raising questions about its origin. The two dominant interpretations favor hot exozodiacal dust or a faint companion, both with the capacity to produce similar infrared interferometric signatures. Aims . We aim to investigate the interferometric signatures of faint companions in the presence of a limb-darkened star and assess their detectability based on visibility and closure phase measurements. By modeling the VLTI instruments and telescope configurations, we explore the limitations of current detection methods and evaluate the challenges in distinguishing between hot exozodiacal dust and a faint companion as the source of the observed infrared excess. Methods . We modeled a system consisting of a limb-darkened star and a faint companion within a field of view of 2 au × 2 au, corresponding to an angular separation of up to 0.07 as from the star. We calculated the visibility and closure phases for three VLTI instruments (PIONIER, GRAVITY, and MATISSE), along with four telescope configurations (small, medium, large, and extended). Results . We derived an upper limit for the companion-induced visibility deficit of |Δ V ( f )| ≤2 f for a companion-to-star flux ratios of f ≤10%, as well as upper limits on the closure phase, which changes linearly with f and is inversely proportional to the visibility of the star. Contrary to the common interpretation that near-zero closure phases rule out the presence of a companion, we show that companions can remain undetected in closure phase data, as indicated by significant non-detection probabilities. However, these companions can still produce measurable visibility deficits that can even approach the theoretical upper limit. We confirmed our results by reevaluating an L -band observation of κ Tuc A using MATISSE. We found indications of a faint companion with a flux ratio of 0.7% and an estimated non-detection probability of around ∼ 21%, which could explain the variability of the previously observed visibility deficit. Conclusions . Previously applied companion rejection criteria, such as near-zero closure phases and flux estimates based on Gaussiandistributed dust densities, are not universally valid. This highlights the need for a reevaluation of companion rejections in former studies of the hot exozodiacal dust phenomenon. In addition, we propose a novel method for distinguishing both sources of the visibility deficit.