A Systematic Search for Big Dippers in ASAS-SN

Dipper stars are extrinsically variable stars with deep dimming events due to extended, often dusty, structures produced by a wide range of mechanisms such as collisions, protoplanetary evolution or stellar winds. ASAS-SN has discovered 12 dipper-like objects as part of its normal operations. Here we systematically search the $\sim 5.1$ million ASAS-SN targets with $13<g<14$~mag for dippers with $Δg\ge0.3$~mag to identify 4 new candidates. We also discover 15 long-period eclipsing binary candidates. We characterized the 19 new and 12 previously discovered objects using the ASAS-SN light curves and archival multi-wavelength data. We divide them into three categories: long-period eclipsing binaries with a single eclipse (13 total), long-period eclipsing binaries with multiple eclipses (7 total) and dipper stars with dust or disk occultations (11 total).

💡 Research Summary

This paper presents a systematic search for “big dipper” stars—objects that exhibit deep (Δg ≥ 0.3 mag), long‑lasting, non‑periodic dimming events—in the All‑Sky Automated Survey for SuperNovae (ASAS‑SN) database. The authors focus on the magnitude range 13 < g < 14 mag, which provides a large sample of unsaturated, high‑signal‑to‑noise light curves. Using the SkyPatrol v2 pipeline, they retrieve g‑band light curves for ~5.1 million sources and apply a three‑step filtering process: (1) identify any point with a depth ≥ 0.3 mag, (2) discard high‑amplitude continuous variables (σ > 0.15 mag) to remove short‑period eclipsing binaries, YSOs, and R Coronae Borealis stars, and (3) require that a dip be seen in at least two independent cameras within a 1‑day window to suppress instrumental artifacts. This reduces the candidate list to ~11 000 objects, of which ~1 400 are known variables. Visual inspection eliminates systematic false positives (bleeding from bright stars, camera‑specific glitches, polar field‑rotation issues) and astrophysical contaminants, leaving 19 new candidates plus 11 of the 12 previously known dipper‑like objects (one missed because its dip was shallower than the threshold).

The final sample is divided into three categories based on light‑curve morphology and ancillary data: (i) single‑dip eclipsing binaries (SDE) with one symmetric eclipse, implying periods longer than the ~10‑year ASAS‑SN baseline; (ii) multi‑dip eclipsing binaries (MDE) showing multiple eclipses, allowing period estimation via Lomb‑Scargle analysis and Gaia radial‑velocity amplitude (RV_amp) and renormalized unit weight error (RUWE) diagnostics; and (iii) big dipper stars, characterized by asymmetric, long‑duration dimming events likely caused by occultations from dusty structures (e.g., debris clouds, circumstellar disks, wind‑driven dust tails).

For the SDE group, the authors discuss individual objects such as ASASSN‑23ht (Δg ≈ 1.2 mag, 114‑day duration) and ASASSN‑24cf (Δg ≈ 2.3 mag, 95‑day duration), noting that the depth‑duration combinations are at the extreme end of what is plausible for eclipsing binaries, raising the possibility of true dipper phenomena. Period lower limits are derived by folding the light curves and considering observational gaps, often yielding periods ≥ 8 years.

In the MDE group, multiple eclipses enable period determination. Gaia RV_amp > 20 km s⁻¹ and RUWE > 1.4 are used as robust binary indicators. For example, J223332+565552 shows an RV_amp of 22.29 km s⁻¹, leading to an estimated orbital period of ~873 days under the assumption of an edge‑on circular orbit.

The big dipper candidates are examined with multi‑wavelength spectral energy distributions (SEDs) built from WISE, 2MASS, GALEX, and Gaia photometry. Mid‑infrared excesses are detected in most cases, supporting the presence of circumstellar dust. The authors discuss plausible physical mechanisms: (a) dust clouds generated by collisions of planetesimals, (b) occultations by warped or precessing protoplanetary disks, and (c) large dust tails driven by stellar winds. One object (J114712‑621037) is omitted from the primary sample because its dip depth falls below the 0.3 mag threshold despite a ~1000‑day duration, illustrating the trade‑off between completeness and manageable false‑positive rates.

Methodologically, the paper highlights the utility of combining time‑domain photometry with Gaia astrometric and spectroscopic diagnostics. The RUWE metric flags astrometric excess noise indicative of unresolved binaries, while RV_amp provides a direct probe of orbital motion. This dual approach efficiently separates genuine eclipsing binaries from other variable phenomena.

The authors conclude by outlining future extensions: lowering the Δg threshold, incorporating longer baseline data (e.g., Gaia DR4, LSST), and applying machine‑learning classifiers trained on the multi‑camera, multi‑band features identified here. Such enhancements would increase sensitivity to shallower, longer‑period events and enable statistical studies of the incidence of dusty occultations across stellar populations, thereby informing models of stellar evolution, disk dynamics, and planet formation.