🔭 astrophysics

The eROSITA Final Equatorial Depth Survey (eFEDS): SDSS spectroscopic observations of X-ray sources

This paper presents a comprehensive optical spectroscopic survey of 13,079 X-ray selected sources in the eFEDS field using SDSS, confirming that the sample is predominantly composed of diverse active galactic nuclei (AGNs) with reliable redshifts to support the upcoming Black Hole Mapper program.

Original authors: Catarina Aydar, Andrea Merloni, Tom Dwelly, Johan Comparat, Mara Salvato, Johannes Buchner, Marcella Brusa, Teng Liu, Julien Wolf, Scott F. Anderson, Carolina P. Andonie, Franz Erik Bauer, Michael R.

Published 2026-06-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Catarina Aydar, Andrea Merloni, Tom Dwelly, Johan Comparat, Mara Salvato, Johannes Buchner, Marcella Brusa, Teng Liu, Julien Wolf, Scott F. Anderson, Carolina P. Andonie, Franz Erik Bauer, Michael R. Blanton, William Nielsen Brandt, Yaherlyn Díaz, Lorena Hernandez-García, Dong-Woo Kim, Takamitsu Miyaji, Sean Morrison, Blessing Musiimenta, Castalia Alenka Negrete, Qingling Ni, Claudio Ricci, Donald P. Schneider, Axel Schwope, Yue Shen, Sophia G. H. Waddell, Riccardo Arcodia, Dmitry Bizyaev, Joseph N. Burchett, Priyanka Chakraborty, Kevin Covey, Boris T. Gansicke, Antonis Georgakakis, Paul J. Green, Hector Ibarra, Jacob Ider-Chitham, Anton M. Koekemoer, Juna A. Kollmeier, Mirko Krumpe, Georg Lamer, Adam Malyali, Kirpal Nandra, Kaike Pan, Claudio Rivera Pizarro, José Sanchez-Gallego, Jonathan R. Trump, Tanya Urrutia

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe as a giant, dark ocean. For a long time, astronomers could only see the "islands" (stars and galaxies) that glowed in visible light, like lighthouses. But many of the most energetic, violent events in the universe—like supermassive black holes eating matter—don't shine brightly in visible light. Instead, they scream in X-rays, a form of light our eyes can't see.

This paper is about a massive project called eFEDS (the eROSITA Final Equatorial Depth Survey) that acted like a deep-sea sonar, scanning a specific patch of the sky to find these hidden X-ray "screams." However, finding the scream isn't enough; to understand who is screaming and how far away they are, astronomers need to look at them with "normal" eyes (optical telescopes) and take their "ID cards" (spectra).

Here is the story of how they did it, explained simply:

1. The Great X-Ray Hunt

The team used a powerful X-ray telescope called eROSITA (riding on a satellite named SRG) to scan the sky. It found nearly 28,000 point-like sources (tiny dots of X-ray light) in a specific area of the sky called the eFEDS field.

The Problem: X-ray telescopes are great at finding where something is, but they are terrible at telling you what it is or how far away it is. It's like hearing a siren in the fog; you know something is there, but you don't know if it's an ambulance, a fire truck, or a police car, or how far down the road it is.

2. The Optical "ID Check"

To solve this, the team turned to the Sloan Digital Sky Survey (SDSS), a giant optical telescope that acts like a massive library of star and galaxy "ID cards."

The Analogy: Imagine you have a list of 28,000 mysterious phone numbers (the X-ray sources). You need to call them to get their names and addresses. The SDSS telescope is the phone book.
The Challenge: The team had to match the X-ray "phone numbers" to the optical "names" in the library. They gathered data from many different surveys (like GAMA, WiggleZ, and LAMOST) to build the most complete list possible.

3. The Human "Proofreaders"

Computers are great at reading these ID cards, but they sometimes make mistakes, especially when the signal is faint or the object is weird.

The Process: The team didn't just trust the computer. They hired a team of human "proofreaders" to visually inspect over 14,000 spectra (the detailed light fingerprints of the objects).
The Result: This human check was crucial. It fixed computer errors and confirmed that 12,000 of these objects had reliable "ID cards" (redshifts, which tell us distance). They found that 99% of the objects with a clear enough signal could be identified correctly.

4. What Did They Find?

Once they had the names and addresses, they looked at the "demographics" of the crowd:

The Party Guests: The vast majority (about 97%) of these X-ray sources turned out to be Active Galactic Nuclei (AGNs). These are supermassive black holes at the centers of galaxies that are actively eating gas and dust, glowing brightly in X-rays.
The Locals: Only about 3% were "locals" (objects inside our own Milky Way galaxy, like stars or stellar remnants).
The Diversity: The team found a huge variety of these black holes. Some were "blue" and bright (unobscured, easy to see), while others were "red" and dimmer (hidden behind dust and gas).

5. The "Stacking" Trick

To see the faint details of these objects, the astronomers used a technique called stacking.

The Analogy: Imagine trying to hear a whisper in a noisy room. If you listen to one person, you might miss it. But if you record 1,000 people whispering the same thing at the same time and average the recordings, the whisper becomes loud and clear.
The Application: They grouped thousands of similar galaxies together (based on their color and distance) and averaged their light spectra. This allowed them to see faint details, like specific chemical elements and the shape of the black hole's "voice," which would be invisible in a single object.

6. The Big Picture

This paper is essentially a pilot study (a test run).

The Goal: It proved that the method works. They successfully matched X-ray sources to optical spectra and cleaned up the data.
The Future: This success paves the way for a much bigger project called the Black Hole Mapper (BHM). The goal is to repeat this process for hundreds of thousands of X-ray sources across the entire sky, creating a massive, uniform map of how black holes have grown and evolved over the history of the universe.

In summary: The team used a high-tech X-ray sonar to find hidden black holes, then used a giant optical library and a team of human proofreaders to identify them. They found that the universe is full of hungry black holes, and they have now built the perfect toolkit to study them in even greater detail in the future.

Technical Summary: The eFEDS SDSS Spectroscopic Follow-up

Problem and Motivation
To understand the evolution of astrophysical populations, particularly Active Galactic Nuclei (AGNs), large-scale statistical samples are required. While the SRG/eROSITA all-sky survey has detected nearly one million X-ray sources, interpreting X-ray spectra and deriving physical properties (such as luminosity and black hole mass) requires accurate redshifts. Optical spectroscopy provides these redshifts and classifies sources. However, previous X-ray surveys (e.g., ROSAT, XMM-Newton, Chandra) were limited by small fields of view or heterogeneous observation settings, making uniform statistical analysis difficult. The eROSITA Final Equatorial Depth Survey (eFEDS) serves as a performance verification field, offering a deep, contiguous X-ray dataset (~140 deg²) that is ~40% deeper than the planned all-sky survey. The challenge addressed by this paper is the systematic acquisition, validation, and compilation of optical spectroscopic redshifts for the eFEDS X-ray sources to create a reliable, uniform catalogue for demographic studies.

Methodology
The authors constructed a value-added catalogue by integrating X-ray data from eROSITA with optical spectroscopic data from the Sloan Digital Sky Survey (SDSS) and other archival surveys.

Data Compilation:
- X-ray Source: The sample is drawn from the eFEDS Main catalogue (Brunner et al. 2022), containing 27,369 point-like X-ray sources.
- Spectroscopic Redshifts (Spec-z): The authors aggregated redshifts from multiple sources:
  - SDSS: Including legacy surveys (SDSS-I/II/III), the BOSS survey, and dedicated SDSS-IV and SDSS-V plate programs specifically targeting eFEDS counterparts.
  - External Surveys: GAMA, WiggleZ, 2SLAQ, 6dFGS, LAMOST, Gaia RVS, FAST, Hectospec, 2MRS, and archival data from SIMBAD and NED.
- Matching: A dedicated Python package (efeds_speccomp) was used to merge these heterogeneous catalogues. The process involved filtering by quality, normalizing quality grades (NORMQ), and associating spectra with unique astrophysical objects defined in the DESI Legacy Imaging Survey (DR9).
Visual Inspection (VI) and Validation:
- A critical step involved the visual inspection of 14,895 SDSS spectra covering 13,079 unique sources.
- Goal: To verify the reliability of the SDSS pipeline redshifts (which can fail for certain object types) and assign classifications.
- Process: A team of volunteers inspected spectra using an interactive web tool, comparing observed spectra against pipeline models. They assigned a normalized redshift quality metric (NORMQ):
  - NORMQ = 3: Secure redshift (confirmed by VI or high-confidence pipeline).
  - NORMQ = 2: Lower confidence redshift (e.g., low S/N, single line).
  - NORMQ = 1: Spectrum available but no useful redshift.
  - NORMQ = -1: Blazar-like (no redshift).
- This process allowed the recovery of reliable redshifts for sources where the pipeline failed or flagged warnings.
Sample Selection for Analysis:
- The final spectroscopic sample consists of 12,011 sources with reliable redshifts (NORMQ = 3).
- For spectral stacking analysis, a pure AGN sample was defined by applying cuts: $L_{0.2-2.3 \text{ keV}} > 10^{42}$ erg s $^{-1}$ , $CLUSTER\_CLASS < 4$ (to exclude cluster members), and $z > 0.001$ (to exclude Galactic objects), resulting in 9,660 AGN candidates.

Key Results

Catalogue Statistics: The study provides spectroscopic redshifts for 13,079 eFEDS sources (48% of the total point-like sample). Of these, 12,011 (44% of the total) have high-quality, reliable redshifts (NORMQ = 3).
Completeness and Purity:
- The spectroscopic completeness increases from 48% for the full sample to 81% for a "cleaner" sample defined by reliable counterparts, high X-ray detection likelihood, and optimal sky coverage ( $r < 21.38$ ).
- Visual inspection recovered reliable redshifts for 99% of spectra with Signal-to-Noise (S/N) > 2.
- The sample is dominated by AGNs; only ~3% of the SDSS spectroscopic sources are Galactic objects.
Source Classification:
- The majority of sources are broad-line AGNs (QSOs).
- The sample includes narrow-line galaxies, optically passive galaxies, and a small fraction of BAL QSOs and Blazars.
- The X-ray selection effectively detects reddened AGNs that might be missed by optical color selection alone.
Spectral Stacking Analysis:
- The authors created high S/N spectral stacks for AGNs binned by redshift ( $z < 0.5$ , $0.5 \le z < 1.0$ , $1.0 \le z < 2.0$ , $z \ge 2.0$ ) and optical color ( $g-r$ ).
- Low-z ( $z < 0.5$ ): Redder stacks are dominated by host galaxy stellar populations (showing absorption lines like Ca H&K and Na I D), while bluer stacks are quasar-dominated (showing strong emission lines like Mg II and H $\beta$ ).
- High-z ( $z \ge 2$ ): Stacks are predominantly quasar-dominated, showing broad emission lines (C IV, Mg II) consistent with Vanden Berk et al. (2001) templates, though X-ray selection captures redder quasars.
- Obstruction Trends: Redder objects generally show higher median X-ray column densities ( $N_H$ ), suggesting a correlation between optical color and obscuration, though this trend varies with redshift.

Significance and Claims
The paper claims to present one of the largest uniform optical spectroscopic surveys of X-ray-selected sources to date. Its primary significance lies in:

Validation: Demonstrating that the SDSS pipeline is highly efficient for X-ray selected sources but requires visual inspection to recover ~2% of catastrophic failures and to assign robust classifications.
Demographic Utility: Providing a large, uniform sample that enables statistical studies of AGN evolution, black hole growth, and the role of obscuration across cosmic time.
Pathfinding: Serving as a pilot study for the Black Hole Mapper (BHM) program of SDSS-V, which aims to obtain optical spectra for hundreds of thousands of eROSITA sources. The data models and visual inspection results are intended to train machine-learning pipelines for future, larger data releases.
Diversity: Highlighting that X-ray selection reveals a diverse population of AGNs, including obscured and reddened quasars, which complements optical-only selection methods.

The authors emphasize that while the stacking analysis provides qualitative trends, quantitative analysis of individual spectral features (e.g., black hole mass estimation, outflow detection) is reserved for future work. The catalogue itself is made available to the community to facilitate further statistical studies.