A speckle search for the outer companion of KIC 9832227

We present Gemini-N/'Alopeke speckle observations of KIC 9832227, a system originally predicted to become a red nova. The diffraction limited observations do not find an outer companion between 11 and 678 AU that could be responsible for the period changes of the system.


INTRODUCTION
KIC 9832227 is a contact binary with a changing orbital period which was originally thought to be exponentially declining, predicting a merger as a red nova by 2022 (Molnar et al. 2017). This prediction was later negated based on a data re-analysis (Socia et al. 2018), but the origin of the change in the period has remained unknown, with an outer companion on a hierarchical triplet being the most likely culprit (Kovacs et al. 2019). Period changes in contact binaries, both positive and negative, are not unheard of. One illustrative example is the famous contact binary star W UMa whose changing period was well documented in the 1970's (e.g. Cester & Pucillo 1972;Cester et al. 1976).
In this Research Note we exploit diffraction limit observations with the Gemini North 8.1m telescope to probe the existence of outer companions to KIC 9832227.

OBSERVATIONS
Observations were carried out with the speckle interferometer 'Alopeke (Scott et al. 2021), mounted at the Gemini North Telescope in MaunaKea. 20×1000×60ms exposures in medium band filters centered at 562nm and 832nm were obtained simultaneously under Gemini program GN-2022A-LP-205 on May 10, 2022. Data reduction of the speckle images is explained in Howell et al. (2011).
The speckle observations detected no companions around KIC 9832227 between 20 mas and 1.2 arcsec, above the 5-σ contrast curves shown in Fig. 1 (left panel). These angular limits equal a 11.3 AU to 678 AU range when using the distance to the system derived by Molnar et al. (2017) of 565 pc. Gaia DR3 gives a slightly farther distance, distance_gspphot=619.92 pc (Gaia Collaboration et al. 2022), but given the large astrometric noise of the observation we use the Molnar et al. (2017) distance in the following. The relative magnitude contrast curves were calibrated into absolute magnitudes using the Gaia magnitude for the system (G = 12.2), a G0V spectrum from the Pickles (1998) library (consistent with the temperature of 5 828 K measured by Molnar et al. (2017)) and the 'Alopeke transmission filter curves 1 using pyphot 2 .
To transform the contrast curves into mass limits for any companion we used the MIST stellar models (Choi et al. 2016) transformed into the 'Alopeke filters using SPISEA (Hosek et al. 2020, Fig. 1, right panel) 3 . At a distance of 100 mas, the 5-σ detection limit in the 832nm filter equals 17.27 mag which transforms into a mass of 0.37 M ; at 1 the detection limit of 20.02 mag yields a mass 0.13 M , irrespective of the age of the system and for a metallicity of [Fe/H]=-0.04 (Molnar et al. 2017).
Even though these observations cannot probe the most up-to-date prediction for the position of the third member of this putative hierarchical system (P ∼13.5 years; ∼ 8.2 AU, Kovacs et al. 2019), they can rule out a possible non-degenerate companion at a larger distance, for example, the component D in the original model of Molnar et al. (2017), expected to have a period longer than 25 years.