[论文解读] Gaia Data Release 3. The first Gaia catalogue of eclipsing binary candidates

N. Mowlavi, B. Holl+12 more|arXiv (Cornell University)|Nov 2, 2022

Stellar, planetary, and galactic studies被引用 5

一句话总结

本文发布了首个基于盖亚数据释放3的食双星候选者星表，包含2,184,477个天体，其G星等从几等至20等，覆盖全天区。该星表利用清洗后的G波段光变曲线进行贝叶斯模型比较，以确定轨道周期，并拟合双高斯加余弦模型，实现93%的周期与文献值一致，相对于OGLE4星表的完整度估计为25–50%，是迄今最大的此类星表。

ABSTRACT

We present the first Gaia catalogue of eclipsing binary candidates released in Gaia DR3, describe its content, provide tips for its usage, estimate its quality, and show illustrative samples. The catalogue contains 2,184,477 sources with G magnitudes up to 20 mag. Candidate selection is based on the results of variable object classification performed within the Gaia Data Processing and Analysis Consortium, further filtered using eclipsing binary-tailored criteria based on the G light curves. To find the orbital period, a large ensemble of trial periods is first acquired using three distinct period search methods applied to the cleaned G light curve. The G light curve is then modelled with up-to two Gaussians and a cosine for each trial period. The best combination of orbital period and geometric model is finally selected using Bayesian model comparison based on the BIC. A global ranking metric is provided to rank the quality of the chosen model between sources. The catalogue is restricted to orbital periods larger than 0.2 days. About 530,000 of the candidates are classified as eclipsing binaries in the literature as well, out of ~600,000 available crossmatches, and 93% of them have published periods compatible with the Gaia periods. Catalogue completeness is estimated to be between 25% and 50%, depending on the sky region, relative to the OGLE4 catalogues of eclipsing binaries towards the Galactic Bulge and the Magellanic Clouds. The analysis of an illustrative sample of ~400,000 candidates with significant parallaxes shows properties in the observational HR diagram as expected for eclipsing binaries. The subsequent analysis of a sub-sample of detached bright candidates provides further hints for the exploitation of the catalogue. The orbital periods, light curve model parameters, and global rankings are all published in the catalogue with their related uncertainties where applicable.

研究动机与目标

• 创建首个基于盖亚数据释放3的全面、全天区食双星候选者星表。
• 开发并应用一种稳健的方法，利用盖亚的多历元测光数据识别食双星，重点关注G波段光变曲线。
• 通过与现有文献交叉匹配并比较轨道周期，估计星表的完整度与可靠性。
• 为每个候选者提供全局排序的模型质量指标，以实现对高保真度系统的优先筛选。
• 通过示范性样本展示星表的实用性，包括分离开双星和周期-偏心率分析。

提出的方法

• 候选者的选择基于盖亚DPAC流水线的可变天体分类结果，随后通过针对食双星特性的G波段光变曲线标准进行筛选。
• 利用三种不同的周期搜索方法对清洗后的G波段光变曲线计算大量试验周期。
• 光变曲线使用每个试验周期最多两个高斯函数和一个余弦函数进行建模，以表示食变与椭球形变亮的形态。
• 通过贝叶斯信息准则（BIC）进行贝叶斯模型比较，选择最优模型，以在拟合质量与模型复杂度之间取得平衡。
• 计算全局排序指标以评估所选模型的整体质量，从而实现对高置信度候选者的优先排序。
• 所有周期大于0.2天的源均发布其轨道周期、光变曲线模型参数及不确定性。

实验结果

研究问题

• RQ1 盖亚DR3食双星星表相对于现有巡天（如OGLE4）的总体完整度如何？
• RQ2 盖亚推导的轨道周期与已知食双星文献值的匹配程度如何？
• RQ3 双高斯加余弦模型在多大程度上能准确表示盖亚DR3数据中食双星的光变曲线形态？
• RQ4 候选者中有多少是新发现的？这一比例在不同天区是否有所不同？
• RQ5 候选者的观测特性（如周期分布和赫罗图中的位置）与食双星的预期特征相比如何？

主要发现

• 星表包含2,184,477个食双星候选者，是迄今发布过的最大此类星表。
• 约530,000个候选者在文献中存在交叉匹配，其中88%在文献中被分类为食双星。
• 88%的交叉匹配源的文献周期与盖亚周期相容，包括考虑周期比为2的情况。
• 基于与OGLE4星表在银河系核球和大、小麦哲伦云区域的比较，星表完整度估计为25–50%，具体取决于天区。
• 具有良好视差的示范样本在赫罗图中表现出预期的分布，证实了样本的物理一致性。
• 对分离开双星的分析揭示了在周期-偏心率空间中的挑战，凸显了仔细后续观测的必要性，也展示了星表在详细研究中的实用性。