Close-in faint companions mimicking interferometric hot exozodiacal dust observations

Context: Interferometric observations of various nearby main-sequence stars show an unexpected infrared excess, raising the question of its origin. The two dominant interpretations favor hot exozodiacal dust or a faint companion, as both can produce similar signatures. Method: We modeled a system consisting of a star and a faint companion within a field of view of 2au x 2au. We calculated the visibility and closure phases for three VLTI instruments (PIONIER, GRAVITY, and MATISSE) and four telescope configurations. Aim: We aim to investigate the interferometric signatures of faint companions and assess their detectability. We explore 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 excess. Results: We derived an upper limit for the companion-induced visibility deficit and closure phase. 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, yet, these companions can still produce measurable visibility deficits. We confirmed our results by reevaluating an L-band observation of kappa Tuc A. We found indications for 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: Previous companion rejection criteria, such as near-zero closure phases and flux estimates based on Gaussian-distributed 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 method for distinguishing both sources of visibility deficit.

💡 Research Summary

The paper investigates whether faint, close‑in companions can masquerade as hot exozodiacal dust in infrared interferometric observations. Using a model that combines a limb‑darkened stellar photosphere with a point‑like companion located within a 2 au × 2 au field of view (≈0.07″ at the distance of the target stars), the authors compute complex visibilities and closure phases for three VLTI instruments—PIONIER (H band), GRAVITY (K band), and MATISSE (L band)—and for four baseline configurations (small, medium, large, extended).

Analytically, the total complex visibility of the system is Vₛ₊c = (Vₛ + V_c f)/(1 + f) where f = f_c/f_s is the companion‑to‑star flux ratio. Expanding for f ≪ 1 and assuming the stellar visibility Vₛ≈1, they obtain linear approximations: Vₛ₊c ≈ Vₛ + (V_c − Vₛ) f and the visibility deficit ΔV = Vₛ − Vₛ₊c ≈ (Vₛ − Re V_c) f. Because the real part of the companion’s visibility can be as low as –1, the absolute deficit is bounded by |ΔV| ≤ 2 f. Thus a companion contributing 1 % of the total flux can produce up to a 2 % drop in fringe contrast.

Closure phase Φ, defined as the sum of three baseline phases, is highly sensitive to point‑symmetry breaking. While a symmetric dust ring yields Φ≈0°, a binary produces a non‑zero phase ψ_c that depends on the companion’s position (x,y) relative to the baseline triangle. However, for certain geometries—especially when the companion lies near the symmetry axis of the baseline triangle—the phase contribution can cancel, resulting in Φ values indistinguishable from zero within typical measurement uncertainties. The authors quantify this effect by calculating a “non‑detection probability” that a companion of given f and separation would produce a closure phase below the detection threshold. For f = 0.7 % at a projected separation of ≈1.4 au, the probability is ≈21 %.

Numerical simulations incorporate realistic instrumental noise, spectral resolution, and baseline lengths for each instrument. The results show that visibility deficits scale linearly with f across all bands, while closure phase sensitivity improves at shorter wavelengths (H, K) and for longer baselines. Nevertheless, even the most favorable configurations cannot guarantee companion detection when the phase signal is suppressed.

The paper applies this framework to κ Tuc A (HD 7788), a well‑studied F6 IV‑V star at 21 pc that exhibits variable near‑infrared excess. Previous H‑band PIONIER data reported visibility deficits of 1.43 % (2012) and 1.16 % (2014) with near‑zero closure phases, leading to the assumption that hot dust was responsible. L‑band MATISSE observations later inferred a dust‑to‑star flux ratio of 5–7 % from a narrow ring at 0.1–0.29 au. Re‑analysing the L‑band data with the binary model, the authors find that a companion with f ≈ 0.7 % reproduces the observed visibility deficit while yielding a closure phase that would be missed in ≈21 % of trials. This demonstrates that the standard “near‑zero closure phase ⇒ no companion” criterion is not universally valid.

Finally, the authors propose a diagnostic that combines visibility amplitude and closure phase simultaneously. Because a companion induces a correlated change—visibility deficit proportional to f and a phase shift that scales with the inverse of the stellar visibility—while a symmetric dust ring produces a deficit with essentially zero phase, plotting ΔV versus Φ for a given dataset can statistically separate the two scenarios. This method, together with multi‑band observations and diverse baseline configurations, offers a robust pathway to discriminate between hot exozodiacal dust and faint companions.

In summary, the study provides (1) an analytic upper bound for companion‑induced visibility deficits, (2) a quantitative assessment of closure‑phase non‑detectability, (3) a thorough simulation of instrument‑specific sensitivities, (4) a re‑interpretation of κ Tuc A data that challenges previous dust‑only explanations, and (5) a practical combined‑observable strategy for future interferometric surveys. These results call for a reassessment of past hot‑dust detections and highlight the need for careful companion‑exclusion analyses in high‑precision infrared interferometry.