Quasar identifications from the slitless spectra: a test from 3D-HST

Yuxuan Pang, Xue-Bing Wu, Yuming Fu, Rui Zhu, Tao Ji, Qinchun Ma, Xiaotong Feng

First listed 2025-05-20 | Last updated 2025-05-20

Abstract

Slitless spectroscopy is a traditional method for selecting quasars. In this paper, we develop a procedure for selecting quasars (QSOs) using the 3D-HST G141 slitless spectra. We initially identify over 6,000 sources with emission lines broader than those typically found in emission line galaxies (ELGs) by analyzing the 1D spectra. These ``broad'' emission lines may originate from actual QSO broad lines ($\rm FWHM\geq1200~\rm km/s$) or the convolved narrow lines ($\rm FWHM = 200\sim 300\rm km/s$) in ELGs with effective radii $\geq$0.3" (2.5Kpc at z=1). We then propose a criterion based on the reliability of the QSO component in the forward modeling results. Using the known QSOs, ELGs, and simulation spectra, we find that our criterion successfully selects about 90\% of known QSOs with H$α$ or H$β$ line detection and point-like structures, with an ELG contamination rate of about 5\%. We apply this method to emission line sources without significant contamination and select 19 QSO candidates at redshift $z=0.12-1.56$. 12 of these candidates have Chandra X-ray detections. This sample covers a broader range of the rest-frame UV colors and has redder optical slopes compared to the SDSS QSOs, yet it is more likely to be composed of normal QSOs rather than little red dots. Through spectral analysis, the estimated black hole masses of the sample are $10^{6.9}-10^{8.3} M_{\odot}$. Our new candidates improve the completeness of the QSO sample at $z=0.8-1.6$ in the 3D-HST field. The proposed method will also be helpful for QSO selections via slitless spectroscopy in Euclid and the Chinese Space Station Telescope.

Short digest

Pang et al. build a slitless-spectrum QSO selector on 3D-HST/G141 by first flagging >6,000 sources with apparently broad lines and then using forward modeling to test whether the “broad” features require a point-like QSO component rather than size-broadened ELG lines. Calibrated on known QSOs, ELGs, and simulations, the criterion recovers ~90% of point-like QSOs with Hα/Hβ while keeping ELG contamination at ~5%. Applied to clean emission-line sources, it yields 19 QSO candidates at z=0.12–1.56 (12 X-ray detections) with broader UV colors and redder optical slopes than SDSS QSOs and MBH≈10^6.9–10^8.3 M⊙, consistent with normal QSOs rather than little red dots. The method boosts QSO completeness at z≈0.8–1.6 in 3D-HST and is directly portable to Euclid and CSST slitless surveys.

Key figures to inspect

• Figure 1 — Walk through the two-step pipeline; focus on how the forward-model “QSO component reliability” gate is defined and where size-broadened ELG lines fail the test.
• Figure 2 — Inspect the continuum-fit and thresholding on 1D spectra; compare the distorted versus clean examples to see how peaks are flagged and how apparent FWHM≥1200 km/s are distinguished from noise.
• Figure 3 — Examine heavy-contamination cases in 2D spectra; compare the data, model, and contamination model to see how spurious broadening from overlaps is recognized and removed.
• Figure 4 — Check the F140W–redshift distribution for the one-line vs two-line subsamples; note that most lie at 0.6–2.4 where Hα or Hβ+[O III] enters G141, setting the method’s most effective redshift window.