Hunting for overlooked eccentric eclipsing binaries from ASAS-3 survey

Highlights

• New non-zero eccentricity eclipsing binaries in ASAS survey.

• Possible tidally-excited oscillations in eclipsing binaries.

• New representation of the light curves for unsupervised machine learning.

Abstract

We present the results of searching for new candidates of eclipsing binaries with eccentric orbits in the ASAS Catalog of Variable Stars (ACVS) using publicly available data from the All Sky Automated Survey (ASAS) and the Transiting Exoplanet Survey Satellite (TESS). Unsupervised machine learning techniques were applied to find anomalies among the light curves of eclipsing binaries. The light curves modeling were performed using JKTEBOP code. The pulsation analysis was done with FAMIAS. We identified 19 new eclipsing binary candidates with non-zero eccentricities in the ACVS, including 10 candidates with eccentricities e ≥ 0.1. Estimates of eccentricities are given. We also report on possible presence of the small-amplitude stellar pulsations at least in two of the reported systems.

Introduction

Binary and multiple systems are ubiquitous among stars (Duquennoy and Mayor, 1991; Fischer and Marcy, 1992; Raghavan et al., 2010; Duchêne and Kraus, 2013; Tokovinin, 2021), especially at birth (Ghez et al., 1993; Leinert et al., 1993; Reipurth and Zinnecker, 1993; Simon et al., 1995). They are the source of information about basic stellar parameters that provide clues to star formation mechanisms, stellar populations, evolution theory, stellar structure models, exoplanet demographics, chemical and dynamical history of the Galaxy (e.g. Pols et al. 1997; White and Ghez 2001; Ciardi et al. 2015; De Marco and Izzard 2017; del Burgo and Allende Prieto 2018; Stanway and Eldridge 2018; Marzari and Thebault 2019). The formation models for multiple systems is still a hot topic and under scrutiny (Tokovinin, 2021).

Only a small fraction of all binaries are eclipsing ones, but they play a crucial role in our understanding of stellar properties which might be derived from the geometry of their orbits. Possibly the most important class of eclipsing system is the detached eclipsing binaries, because their components have not experienced mass transfer, evolve as if they were single stars and so they represent the population of normal stars. From eclipsing binary light curves we can derive the inclination angle and the sum of relative radii of component stars. If the star is also a double-line spectroscopic binary, then the simultaneous analysis of the light curve and the radial velocity curve provides the individual masses, effective temperatures and the length scale of the system. The ephemeris curve can be used to get the apsidal motion period and rate. The rate of apsidal motion depends on the density gradient and thus defines the internal structure of the stars. The mean internal structure constant is based on the observed apsidal motion period corrected for relativistic effects (for example, Claret and Willems 2002) and estimates the degree of central mass concentration. Combining high-quality photometric and spectroscopic observations one can obtain mass and radius of both stars with accuracy up to ∼3% level or better (see e.g. Torres et al. 2010; Southworth 2012) that offers an opportunity to test stellar structure and evolution models (e.g. Lastennet and Valls-Gabaud 2002; Higl and Weiss 2017; Tkachenko et al. 2020). Radii and temperatures of the components are used as distance candles (Paczynski and Sasselov, 1997; Clausen, 2005), to determine the size and structure of the Galaxy (Minniti et al., 2010; Helminiak et al., 2013), distances to the nearby galaxies in the Local Group as well as cosmological distance scale (e.g. Kaluzny et al. 1998; Wyithe and Wilson 2001; Bonanos et al. 2006; Riess et al. 2016; Pietrzyński et al. 2019; Freedman et al. 2020).

An important subgroup amongst the eclipsing binaries is represented by eccentric systems. It is currently believed that multiple stellar systems result from the fragmentation that produces mainly eccentric binaries (Bonnell and Bastien, 1992; Bate and Bonnell, 1997). “Eccentric binaries – still interesting targets” (Zasche, 2018), that help us to study some interesting phenomena. The rate of apsidal motion, that is directly observable, offered the first observational probe of stellar interior structure and evolution (e.g. Pritchard 1997; Claret and Giménez 2010) before asteroseismology, as well as to test various gravitational theories (e.g. Li 2010; De Laurentis et al. 2012) and the possibility of verification of general relativity outside the Solar System (e.g. Gimenez 1985; Wolf et al. 2010). The observed periods of apsidal motion can be used to determine the masses and to estimate the angular axial velocities of components (Benvenuto et al., 2002; Khaliullin and Khaliullina, 2007). Systems that exhibit simultaneously apsidal motion and light-time effects have a third body detected, which makes them even more dynamically interesting objects (Bozkurt and Deǧirmenci, 2007; Borkovits et al., 2015). Eccentricity is also an important diagnostic of binary and multiple systems formations (Kiminki and Kobulnicky, 2012). The orbital and physical parameters of the eccentric systems have been used to constrain the mechanisms and timescales of synchronization of stellar rotation with orbital motion and circularization of orbits predicted by some tidal theories (Meibom and Mathieu, 2005; Devor et al., 2008; Zahn, 2008; Khaliullin and Khaliullina, 2010). High-accuracy observations of the space missions made it possible to detect such subtle effects as tidally induced brightening around the periastron and oscillations, theoretically predicted by Kumar et al. (1995). The interactions of the eccentric stars and their envelopes result in some exotic effects such as tidally trapped pulsations, tidal spin-orbit alignment, orbital decay, Doppler beaming (e.g. Mazeh 2008; Groot 2012; Thompson et al. 2012; Sun et al. 2018; Fuller et al. 2020). Pulsating stars in eclipsing binary systems can provide even more accurate stellar parameters than single pulsators (Grigahcène et al., 2010; Lampens, 2021).

As was noted by Paczynski (1997), all these important tasks require precise calibration, which is possible to do only for bright systems. The relatively bright sources are also needed for follow-up photometric (especially by amateurs) and spectroscopic observations. Besides, studies involving statistical analysis require many objects that become accessible through surveys, especially the all-sky ones.

Currently, many photometric surveys exist (see e.g. Kim et al. 2018) that have various goals and produce huge amounts of data, generate a large number of light curves and detect a lot of new variable stars.

One of such surveys is the All Sky Automated Survey (ASAS) that was inspired by the idea of Paczynski (1997) of continuous surveying and monitoring of objects brighter than 14 magnitude in the entire sky with small-scale low-cost automated telescopes. Started in 1997 (Pojmanski, 1997) ASAS system monitored approximately 20 million stars south of +28 deg in the magnitude range from 8 to 14 mag in V band (and from 7 to 13 in I). One of the main results is the ASAS Catalog of Variable Stars (ACVS) of the southern and part of the northern hemispheres (DEC < +28 deg) (Pojmanski et al., 2005 and references therein) that contains 50,122 variable stars. Preliminary classification of the objects has been performed with an automated algorithm that took into account some properties of the light curves such as period, amplitude, Fourie coefficients, 2MASS colors and IRAS fluxes. Periodic variables were divided into 12 classes (including 11,099 eclipsing binaries) and all others are included into the catalog as MISC. Many objects had multiple, ambiguous or incorrect class designation (Pojmanski et al., 2005). Namely 38,117 ACVS stars, representing 76% of the catalog, are either labeled as MISC, assigned multiple labels, or have low class confidence, and only remaining 24% of stars have confident ACVS labels (Richards et al., 2012). Paczyński et al. (2006) presented the preliminary analysis of 11,076 ASAS eclipsing binaries including 5384 contact (EC), 2949 semi-detached (ESD), and 2743 detached (ED) systems. A number of articles are devoted to the analysis of the ASAS data. Here we will only review some of them that are relevant to the topic of this article. Richards et al. (2012) released the Machine-learned ASAS Classification Catalog (MACC) that is a probabilistic classification catalog of 50,124 ASAS sources from the ACVS classified into 28 classes. The MACC also gives an anomaly score for each ASAS source, which describes its proximity to objects in the training set. Shivvers et al. (2014) used 2454 β Persei listed in MACC, with the cuts on reported anomaly score and class probability, to find 106 highly eccentric (e ≥ 0.1) bright (V < 12 mag) systems, most of which were newly identified. 132 more eccentric eclipsing binaries were selected by Lee (2015) from 2271 ACVS ED binaries light curves. Again, some rejections were done in particular due to limitations and assumptions applied in DEBIL (Devor, 2005) and MECI (Devor and Charbonneau, 2006). Some eccentric binaries in ACVS were reported in other works.

In this work, we present the searching results for eccentric systems using data for all eclipsing binaries publicly available in the ACVS. For this goal, a novel time series representation for clustering analysis was used. The rest of the paper is organized as follows. Section 2 gives the information about used ASAS data and methodology of searching eccentric eclipsing binary systems. Section 3 provides a list of the new candidates of eccentric eclipsing binaries with their eccentricity estimations. Finally, the results are being discussed in Section 4.