[Paper Review] Classifying Be star variability with TESS I: the southern ecliptic

Jonathan Labadie-Bartz, A. C. Carciofi+5 more|arXiv (Cornell University)|Oct 26, 2020

Stellar, planetary, and galactic studies4 citations

TL;DR

This study analyzes TESS photometry of 432 southern ecliptic classical Be stars, classifying their complex variability across timescales from minutes to tens of days. It finds that 98% are variable above noise, with 85% showing groups of closely spaced frequencies, and reveals that mass ejection events correlate strongly with amplitude modulations in frequency groups, while stochastic and high-frequency (p-mode) signals are also prevalent.

ABSTRACT

TESS photometry is analyzed for 432 classical Be stars observed in the first year of the mission. The often complex and diverse variability of each object in this sample is classified to obtain an understanding of the behavior of this class as a population. 98\% of the systems are variable above the noise level, with timescales spanning nearly the entire range of what is accessible with TESS, from tens of minutes to tens of days. The variability seen with TESS is summarized as follows. Nearly every system contains multiple periodic signals in the frequency regime between about 0.5 -- 4 d$^{-1}$. One or more groups of closely-spaced frequencies is the most common feature, present in 85\% of the sample. Among the Be stars with brightening events that are characteristic of mass ejection episodes (17\% of the full sample, or 30\% of early-type stars), all have at least one frequency group, and the majority of these (83\%) show a concurrent temporary amplitude enhancement in one or more frequency groups. About one third of the sample is dominated by low frequency ($f < 0.5$ d$^{-1}$, and often much lower) variability. Stochastic signals are prominent in about 26\% of the sample, with varying degrees of intensity. Higher frequency signals ($6 < f < 15$ d$^{-1}$) are sometimes seen (in 14\% of the sample) and in most cases likely reflect p mode pulsation. In rare cases ($\sim$3\%), even higher frequencies beyond the traditional p mode regime ($f > 15$ d$^{-1}$) are observed.

Motivation & Objective

• To characterize the full range of photometric variability in classical Be stars using high-precision TESS data.
• To classify the diversity of variability patterns across timescales from tens of minutes to tens of days.
• To investigate the connection between periodic frequency groups, stochastic signals, and mass ejection events in Be stars.
• To determine the prevalence and nature of p-mode pulsations and other oscillation modes in the Be star population.
• To assess the role of rotational and pulsational instabilities in shaping the observed variability patterns.

Proposed method

• Analyzing TESS full-cadence light curves for 432 classical Be stars observed in the first year of the mission.
• Applying Fourier analysis to detect periodic signals, identifying groups of closely spaced frequencies and isolated peaks.
• Classifying variability types based on frequency spectrum morphology: multiperiodic, stochastic, low-frequency trends, and high-frequency signals.
• Using cross-correlation between frequency group behavior and brightening events to identify mass ejection episodes.
• Employing standard asteroseismic tools (Period04, VARTOOLS) and astrophysical modeling to interpret signal characteristics.
• Cross-referencing with Gaia, KELT, and archival data to confirm stellar parameters and variability context.

Experimental results

Research questions

• RQ1 What is the overall prevalence and diversity of photometric variability in classical Be stars as observed by TESS?
• RQ2 How common are groups of closely spaced frequencies in the Be star population, and what is their relationship to mass ejection events?
• RQ3 To what extent do stochastic and high-frequency (p-mode) signals contribute to the total variability spectrum?
• RQ4 How do low-frequency variations (<0.5 d⁻¹) and long-term trends influence the observed light curves?
• RQ5 Is there a systematic link between the presence of frequency groups and the occurrence of brightening events indicative of disk ejection?

Key findings

• 98% of the 432 Be stars in the sample show photometric variability above the TESS noise level across timescales from tens of minutes to tens of days.
• 85% of the sample exhibit one or more groups of closely spaced frequencies in the 0.5–4 d⁻¹ range, indicating complex, multiperiodic pulsation.
• Among Be stars showing brightening events (17% of the sample, or 30% of early-type stars), 83% display concurrent amplitude enhancements in one or more frequency groups.
• Approximately 33% of the sample show dominant low-frequency variability (f < 0.5 d⁻¹), often with long-term, aperiodic trends.
• Stochastic variability is prominent in 26% of the sample, with varying intensity, and is often superimposed on periodic signals.
• High-frequency signals (6–15 d⁻¹) are detected in 14% of the sample and are predominantly attributed to p-mode pulsations; about 3% show signals above 15 d⁻¹, indicating potential high-order p-mode or non-radial oscillations.