[Paper Review] What's the nature of sdA stars?

Ingrid Pelisoli, S. O. Kepler+2 more|arXiv (Cornell University)|Oct 18, 2016

Stellar, planetary, and galactic studies3 citations

TL;DR

This paper investigates the nature of sdA stars—white dwarf candidates with sub-main sequence surface gravity (log g ≈ 5.0–6.5) but A-type spectral features—finding that they are likely a cool extension of the Extremely Low Mass (ELM) white dwarf population rather than main sequence A stars. Spectroscopic log g values may be overestimated due to model limitations, and assuming ELM radii reconciles their colors, distances, and velocities with disk populations, though radial velocity variability remains undetected in SDSS subspectra.

ABSTRACT

White dwarfs with log g lower than 7.0 are called Extremely Low Mass white dwarfs (ELMs). They were first found as companions to pulsars, then to other white dwarfs and main sequence stars (The ELM Survey: 2010 to 2016), and can only be formed in interacting binaries in the age of the Universe. In our SDSS DR12 white dwarf catalog (Kepler et al. 2016), we found a few thousand stars in the effective temperature and surface gravity ranges attributed to ELMs. We have called these objects sdAs, alluding to their narrow hydrogen line spectra showing sub-main sequence log g. One possible explanation for the sdAs is that they are ELMs. Increasing the ELMs sample would help constrain the number of close binaries in the Galaxy. Interestingly, if they turn out to be A stars with an overestimated log g, the distance modulus would put these young stars in the Galaxy's halo.

Motivation & Objective

• To resolve the nature of sdA stars—objects with A-type spectra and sub-main sequence log g—found in SDSS DR12.
• To determine whether sdAs are main sequence A stars with overestimated log g or Extremely Low Mass (ELM) white dwarfs.
• To assess whether the observed colors, proper motions, and velocities of sdAs are consistent with ELMs or A stars.
• To evaluate the role of model uncertainties (e.g., metallicity, helium abundance, collisional broadening) in log g determination.

Proposed method

• Spectroscopic fitting of SDSS DR12 spectra (T_eff: 8000–20,000 K, log g: 3.5–8.0) using hydrogen-dominated atmosphere models with solar abundances.
• Photometric analysis using SDSS ugriz and GALEX FUV/NUV magnitudes to correct for extinction and assess colors.
• Color-color diagrams (u-g vs. g-r and fuv-nuv vs. nuv-g) to compare sdAs with known ELMs and A stars.
• Proper motion and velocity estimation using SDSS and APOP catalog data to test disk vs. halo kinematics.
• Comparison of observed radial velocity variations in SDSS subspectra to expected values for ELM binaries (median semi-amplitude ~220 km/s).
• Modeling of collisional line broadening effects in cool hydrogen lines, though not yet fully implemented due to lack of rigorous calculations.

Experimental results

Research questions

• RQ1 Are sdA stars main sequence A stars with overestimated log g, or are they Extremely Low Mass (ELM) white dwarfs?
• RQ2 Why do sdAs have colors consistent with low log g objects despite spectroscopically derived log g > 5.0?
• RQ3 Do the kinematic properties (proper motions, velocities) of sdAs support a disk or halo population, and how does this depend on assumed radius?
• RQ4 Why is radial velocity variability not detected in SDSS subspectra of sdAs, despite expectations for ELM binaries?
• RQ5 To what extent do model uncertainties (e.g., metallicity, helium abundance, collisional broadening) affect log g determination for sdAs?

Key findings

• The sdA sample (1010 objects) shows cooler colors than known ELMs, suggesting a possible cool ELM population not yet detected.
• Despite spectroscopic log g > 5.0, sdAs lie below the log g = 5.0 model line in color-color diagrams, indicating possible overestimation of log g.
• When assuming main sequence radii, sdAs have inconsistent distances and velocities, with many appearing in the halo and showing high velocities inconsistent with known stellar populations.
• When assuming ELM radii (20× smaller), distances and velocities align with thin and thick disk distributions, supporting the ELM hypothesis.
• Lack of radial velocity variability in SDSS subspectra is likely due to low signal-to-noise and short observation times, not absence of binary motion.
• Model uncertainties—especially unaccounted helium abundance and collisional line broadening—likely contribute to log g overestimation, weakening the A-star hypothesis.