Disk Material Inflates Gaia RUWE Values in Single Stars

Binary systems are a common product of star formation, with young populations holding a particular interest given evidence that the binary fraction is highest in the formation environment (Duchêne & Kraus 2013). Obtaining a better understanding of their occurrence rates and properties as a function of stellar age is critical, as binarity can affect many processes that shape the formation and evolution of both stars and planets (Hurley et al. 2002; Marzari & Thebault 2019). However, binaries can be difficult to find and characterize, as the expensive nature of conducting adaptive-optic (AO) surveys (specifically with large telescopes) is limiting. This can be largely mitigated by looking instead toward Gaia, a survey that measures the three-dimensional spatial and velocity distribution of stars (Gaia Collaboration et al. 2016).

This is useful for both wide (ρ > 1'') and close binaries (ρ < 1''). Wide binaries are easily resolved by the Gaia point-spread-function to 1''. Close binaries that are unresolved can be detected through the re-normalized unit weight error (RUWE) of the astrometric solution (Ziegler et al. 2018). Multiple studies have now shown that RUWE values greater than 1.2 signal the presence of a stellar companion within 1'' and with a G-band contrast ratio <5 mag (Rizzuto et al. 2018, Kraus et al. 2022, in preparation). Additionally, this method offers the potential to identify binary populations at many distinct ages, especially as the number of nearby moving groups and associations with well-known ages also increases with the help of Gaia astrometry (e.g., Miret-Roig et al. 2018; Kounkel & Covey 2019; Ujjwal et al. 2020; Wallace et al. 2021).

In this Research Note we assess whether the Gaia RUWE threshold for binary detection, which was developed using field-age stars (τ > 1 Gyr), applies to the youngest stellar populations. There are various possible systematic effects that are common within younger populations, such as the presence of protoplanetary disks (τ ≲ 10 Myr), and we seek to establish how these effects may influence a star's RUWE. Our analysis is motivated by the empirical sensitivity of the RUWE statistic to deviations from a "single-star" light profile, and determining the likelihood of common binary-related issues affecting this statistic. Our primary focus is to further investigate how the presence of disk material alters this profile in a way that inflates the RUWE value, such as via scattered light that is not centered on the star. Additionally, spot checks of various disk bearing sources that are known to be single from AO monitoring have shown inflated RUWE values (e.g., AA Tau; RUWE = 3.33 Gaia Collaboration et al. 2016). To systematically test this new diagnostic at young ages, we assemble a sample of single stars from AO surveys of the Taurus star-forming region (τ ∼ 0–5 Myr; Krolikowski et al. 2021) and compare the RUWE distributions for sub-samples with and without circumstellar disks. From this sample we evaluate the efficacy of the field star RUWE threshold when applied to young, disk-bearing stars, and determine how best to use the RUWE statistic in the search for binary systems in young populations.

We draw our sample from studies that performed AO surveys of star-forming regions in order to look for binary companions, primarily in Taurus and Upper Scorpius. These studies report contrast limits using the F555W, F775W, and F850LP filters from the Hubble Space Telescope (Kraus et al. 2005, 2006), K', K_s , L, and H' filters from the Keck II 10 m and Palomar Hale 200'' telescopes (Kraus et al. 2008, 2011), and the K_p , K_s , and L_p filters from Keck NIRC2 (Cheetham et al. 2015). From these studies we identified 154 targets that are confirmed to be binary or single. To ensure validity of our sample, we only include single stars that cite a contrast limit >1 mag at 10–40 milliarcseconds, reducing the single-star sample to 122.

We then compile the RUWE and WISE W1−W3 values from the Gaia EDR3 and ALLWISE (Cutri et al. 2021) databases, respectively. The data are presented in Figure 1 panel (a). We set a WISE color cut of 1, where stars with W1−W3 > 1 are expected to have a protoplanetary disk. Likewise, stars with W1−W3 < 1 are presumed to not have disks (Pecaut & Mamajek 2013).

Zoom In Zoom Out Reset image size

We create a histogram of the disk-bearing and disk-free samples in order to compare the two populations (Figure 1, panel (b)). To test whether the two samples are drawn form the same parent sample, we conduct a KS test which results in a value of 0.385 and a corresponding p-value of 0.000276. The cumulative distribution is presented in Figure 1, panel (c). This proves that young, disk-bearing single stars have higher RUWE values on average than young, disk-free stars.

We observe that the traditional RUWE cutoff for single stars at 1.2 is too conservative to account for single stars with disks. We calculate the 95th percentile for our disk-bearing and disk-free sample, producing values of 2.5 and 1.6, respectively. Our disk-free RUWE population is broader than the field-age sample used by Bryson et al. (2020), which has a 95th percentile of 1.15. This may be the result of contamination in our disk-free sample from either hidden companions or disk material not probed by the W1−W3 color. Based on our analysis, we suggest a RUWE cutoff value of 2.5 for the identification of binary candidates in disk-bearing sources.

A total of 6 stars were found to have an RUWE > 1.5 yet W1−W3 < 1, signifying a small population of confirmed single stars with inflated RUWEs that cannot be attributed to the presence of a disk or a known companion. These stars are GSC0620800834, RXJ1602.02221, RXJ1606.22036, ScoPMS021, ScoPMS044, and USco160142.6222923. It is worth investigating these stars further, as they hold a potential for possessing a previously undiscovered protoplanetary disk or a binary companion.

This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement.