Mapping the Shores of the Brown Dwarf Desert I.: Upper Scorpius

TL;DR

This survey of 82 young stars in Upper Scorpius used advanced interferometry to detect stellar and substellar companions, revealing a brown dwarf desert and providing new insights into companion frequency and mass functions.

Contribution

First application of nonredundant aperture-mask interferometry in Upper Scorpius to characterize stellar and substellar companions, including a new analysis of the brown dwarf desert.

Findings

Detected 12 new binary companions below traditional detection limits.

Companion frequency similar to field stars, possibly higher among lower-mass stars.

Found a deficit of brown dwarf companions, consistent with the brown dwarf desert.

Abstract

We present the results of a survey for stellar and substellar companions to 82 young stars in the nearby OB association Upper Scorpius. This survey used nonredundant aperture-mask interferometry to achieve typical contrast limits of DeltaK~5-6 at the diffraction limit, revealing 12 new binary companions that lay below the detection limits of traditional high-resolution imaging; we also summarize a complementary snapshot imaging survey that discovered 7 directly resolved companions. The overall frequency of binary companions (~35+5/-4% at separations of 6-435 AU) appears to be equivalent to field stars of similar mass, but companions could be more common among lower-mass stars than for the field. The companion mass function has statistically significant differences compared to several suggested mass functions for the field, and we suggest an alternate log-normal parameterization of the mass-function. Our survey limits encompass the entire brown dwarf mass range, but we only detected a single companion that might be a brown dwarf; this deficit resembles the so-called ``brown dwarf desert'' that has been observed by radial-velocity planet searches. Finally, our survey's deep detection limits extend into the top of the planetary mass function, reaching 8-12 MJup for half of our sample. We have not identified any planetary companions at high confidence (>99.5%), but we have identified four candidate companions at lower confidence (>97.5%) that merit additional followup to confirm or disprove their existence.