A Systematic Search for Changing-Look Quasars in SDSS

TL;DR

This study systematically searches for changing-look quasars in SDSS data, discovering 10 examples with significant spectral and photometric variability, including the highest-redshift cases, and discusses potential physical mechanisms behind these changes.

Contribution

It introduces a new selection method combining photometric and spectroscopic data to identify changing-look quasars, expanding the known sample to higher redshifts and providing insights into their variability timescales.

Findings

Identified 10 changing-look quasars with BEL variability.

Estimated >15% of strongly variable quasars show changing-look features.

Dimming and brightening are unlikely caused by dust extinction alone.

Abstract

We present a systematic search for changing-look quasars based on repeat photometry from SDSS and Pan-STARRS1, along with repeat spectra from SDSS and SDSS-III BOSS. Objects with large, |\Delta g|>1 mag photometric variations in their light curves are selected as candidates to look for changes in broad emission line (BEL) features. Out of a sample of 1011 objects that satisfy our selection criteria and have more than one epoch of spectroscopy, we find 10 examples of quasars that have variable and/or "changing-look" BEL features. Four of our objects have emerging BELs; five have disappearing BELs, and one object shows tentative evidence for having both emerging and disappearing BELs. With redshifts in the range 0.20 < z < 0.63, this sample includes the highest-redshift changing-look quasars discovered to date. We highlight the quasar J102152.34+464515.6 at z = 0.204. Here, not only have the Balmer emission lines strongly diminished in prominence, including H$\beta$ all but disappearing, but the blue continuum $f_{\nu} \propto \nu^{1/3}$ typical of an AGN is also significantly diminished in the second epoch of spectroscopy. Using our selection criteria, we estimate that >15% of strongly variable luminous quasars display changing-look BEL features on rest-frame timescales of 8 to 10 years. Plausible timescales for variable dust extinction are factors of 2-10 too long to explain the dimming and brightening in these sources, and simple dust reddening models cannot reproduce the BEL changes. On the other hand, an advancement such as disk reprocessing is needed if the observed variations are due to accretion rate changes.