Revealing companions to nearby stars with astrometric acceleration

TL;DR

This study uses adaptive optics imaging to directly detect companions to nearby stars with astrometric acceleration, revealing new binaries and challenging previous assumptions about their nature and prevalence.

Contribution

It provides the first direct imaging detections of companions to astrometric binaries and compares these with simulated populations, uncovering unexpected wide binaries and potential spurious accelerations.

Findings

17 sub-arcsecond companions resolved

Approximately 10% of companions may be dark objects

Some acceleration solutions in Hipparcos data are spurious

Abstract

A subset of 51 Hipparcos astrometric binaries among FG dwarfs within 67pc has been surveyed with the NICI adaptive optics system at Gemini-S, directly resolving for the first time 17 sub-arcsecond companions and 7 wider ones. Using these data together with published speckle interferometry of 57 stars, we compare the statistics of resolved astrometric companions with those of a simulated binary population. The fraction of resolved companions is slightly lower than expected from binary statistics. About 10% of astrometric companions could be "dark" (white dwarfs and close pairs of late M-dwarfs). To our surprise, several binaries are found with companions too wide to explain the acceleration. Re-analysis of selected intermediate astrometric data shows that some acceleration solutions in the original Hipparcos catalog are spurious.