eROSITA源的对称物体识别与分类及其AGN内容表征
摘要
准确计量银河演化中的AGN阶段要求获得大规模、干净的AGN样本。现通过SRG/eROSITA可实现。公开数据发布版本1（DR1，2024年1月31日）包含来自西半部银河天顶辐射区的930,203个源。该数据使能够选取大规模AGN样本并发现稀有源。然而，科学回报依赖于对X射线源进行准确表征，需高质量的多波段数据。本文采用贝叶斯NWAY算法和训练的先验知识，对eRASS1源的光学和红外对称物体进行识别与分类，数据来源于Gaia DR3、CatWISE2020和Legacy Survey DR10（LS10）。通过结合光学/红外和X射线属性、基于参考样本训练的机器学习模型对源进行归类为银河或外星系。对于外星系LS10源，采用Circlez计算光学红移。遍历LS10范围后，所有656,614个eROSITA/DR1源至少存在一个可能的光学对称物体；约有57万个为外星系且可能为AGN。其中约有一半是与AllWISE、Gaia和Quaia AGN目录相比的新检测。由于调查浅薄且评估X射线发射概率时可用特征有限，Gaia和CatWISE2020的对称物体可靠性较低。在银河平原，因星源密度增加，关联错误的概率也随之升高。采用保守的可靠性限制后，分别确定约18,000个Gaia和55,000个CatWISE2020的外星系源。我们发布三个高质量的对称物体目录，以及训练和验证数据集，作为该领域的基准。这些数据集具有多种应用价值，特别是为研究人员构建针对完整性和纯度的AGN样本提供支持，加速寻找宇宙最强力源的探索。
引用
@article{arxiv.2509.02842,
title = {Counterpart identification and classification for eRASS1 and characterisation of the AGN content},
author = {M. Salvato and J. Wolf and T. Dwelly and H. Starck and J. Buchner and R. Shirley and A. Merloni and A. Georgakakis and F. Balzer and M. Brusa and A. Rau and S. Freund and D. Lang and T. Liu and G. Lamer and A. Schwope and W. Roster and S. Waddell and M. Scialpi and Z. Igo and M. Kluge and F. Mannucci and S. Tiwari and D. Homan and M. Krumpe and A. Zenteno and D. Hernandez-Lang and J. Comparat and M. Fabricius and J. Snigula and D. Schlegel and B. A. Weaver and R. Zhou and A. Dey and F. Valdes and A. Myers and S. Juneau and H. Winkler and I. Marquez and F. di Mille and S. Ciroi and M. Schramm and D. A. H. Buckley and J. Brink and M. Gromadzki and J. Robrade and K. Nandra},
journal= {arXiv preprint arXiv:2509.02842},
year = {2026}
}
备注
A&A, Paper accepted. The catalogues of LS10, CW2020 Gaia DR3 counterparts and the training samples from 4XMM and Chandra are available via eROSITA web page (https://erosita.mpe.mpg.de/dr1/AllSkySurveyData_dr1/Catalogues_dr1/ ,files 4-12), Zenodo (https://zenodo.org/records/17404798) and Vizier. In Zenodo, the Jupiter notebook for creating samples of AGN is also available