The DESI Single Fiber Lens Search. I. Four Thousand Spectroscopically Selected Galaxy-Galaxy Gravitational Lens Candidates

This paper presents a new catalog of 4,110 galaxy-galaxy gravitational lens candidates, including 3,887 new discoveries, identified by detecting background [O II] emission lines in DESI spectra of foreground luminous red galaxies, which will serve as a valuable dataset for measuring dark matter substructure and enabling time-delay cosmography.

The Gist

Imagine you are standing in a vast, crowded forest at night, holding a single, narrow flashlight beam. You are looking at a specific tree in front of you (a massive, old galaxy). Suddenly, you notice a faint, glowing firefly buzzing behind that tree. Because the tree is so heavy, its gravity acts like a magnifying glass, bending the light from the firefly so that you can see it, even though it's hidden behind the tree.

This is the story of a new scientific paper by the DESI team. They have just found 4,110 of these "fireflies" (background galaxies) that are being magnified by "trees" (foreground galaxies). Here is the breakdown of their discovery in simple terms:

1. The Setup: A Giant Cosmic Net

The researchers used a massive machine called DESI (Dark Energy Spectroscopic Instrument). Think of DESI as a giant, high-tech vacuum cleaner with 5,000 tiny straws (fibers) that can suck up light from 5,000 different galaxies at the exact same time.

They pointed this machine at 5.8 million specific galaxies called LRGs (Luminous Red Galaxies). These are the "trees" in our analogy: huge, heavy, and relatively quiet.

2. The Trick: Finding the Hidden Firefly

Usually, if you look at a galaxy through a telescope, you just see that one galaxy. But DESI is a spectrograph, meaning it doesn't just take a picture; it breaks the light down into a rainbow (a spectrum) to see what chemicals are inside.

The team looked for a specific chemical signature: Oxygen.

• The Foreground Galaxy (The Tree): It has its own light, but it's mostly old stars, so it doesn't have much glowing oxygen.
• The Background Galaxy (The Firefly): These are young, active galaxies full of new stars. They glow brightly with a specific "double" line of oxygen light.

The Discovery: The team found that in 4,110 cases, the light from the "Tree" (foreground) contained a ghostly echo of the "Firefly's" (background) oxygen light. This means the heavy foreground galaxy is acting as a gravitational lens, bending the light from the distant background galaxy right into the same tiny straw (fiber) as the foreground galaxy.

3. The Filter: Cleaning Up the Noise

Finding these signals is like trying to hear a whisper in a hurricane. The data was messy.

• The Problem: Sometimes the "Tree" itself has weird emission lines that look like the "Firefly," or the sky has noise that mimics the signal.
• The Solution: The team used a three-step computer process (like a bouncer at a club):
  1. Phase 1: They subtracted the expected light of the foreground galaxy to see what was left over.
  2. Phase 2: They used a sophisticated math model (MCMC) to check if the leftover light was a real double-line oxygen signal or just a glitch.
  3. Phase 3: They manually checked the "redshift" (distance) to make sure the background object was actually behind the foreground one, not just a trick of the light.

After this cleaning, they were left with 4,110 strong candidates.

4. The Result: A Treasure Trove of New Lenses

Before this paper, astronomers knew of about 400 confirmed gravitational lenses. This new list adds 3,887 new discoveries.

• The Odds: The team calculated that about 53% of these are real lenses (the other half might just be a lucky alignment where two galaxies happen to line up by chance). Even if only half are real, that doubles the number of known lenses in the universe!
• The Size: These lenses are "small" in cosmic terms. The "Einstein Ring" (the circle of light created by the lens) is tiny, often smaller than the width of a human hair seen from a mile away. This is why ground-based cameras can't see them clearly; they need the "magnifying glass" effect of the lens to be detected spectroscopically first.

5. Why Does This Matter?

Why do we care about finding these cosmic magnifying glasses?

• Dark Matter Detective Work: These lenses allow us to map Dark Matter (the invisible stuff that holds galaxies together) in incredible detail. We can see how dark matter is clumped up inside galaxies, which helps us understand if our theories about the universe are correct.
• Time Travel (Sort of): If a supernova (a dying star) explodes behind one of these lenses, the light will take different paths around the galaxy. One path might be longer than the other, so the explosion will appear at different times. By measuring this delay, we can calculate the Hubble Constant—a number that tells us how fast the universe is expanding.
• The Future: The paper mentions that future telescopes like Euclid (a space telescope) and Rubin (a giant ground-based camera) will take high-resolution pictures of these 4,110 candidates to confirm them and watch for those time-delayed supernovae.

Summary Analogy

Imagine you are in a library (the universe) looking for a specific book (a distant galaxy) that is hidden behind a heavy encyclopedia (a foreground galaxy).

• Old Method: You try to take a photo of the encyclopedia, hoping to see the book peeking out. But the photo is too blurry.
• DESI's Method: You shine a light through the encyclopedia. Even though you can't see the book, you can smell the ink (the oxygen light) coming from the book through the encyclopedia.
• The Result: You have now found 4,110 books that were previously invisible, proving that the encyclopedia is acting as a magnifying glass. This helps us understand the weight of the encyclopedia (Dark Matter) and measure the size of the library (the Universe).

This paper is a massive leap forward, turning a handful of known cosmic lenses into a library of thousands, ready for the next generation of telescopes to explore.

Technical Summary

Here is a detailed technical summary of the paper "The DESI Single Fiber Lens Search. I. Four Thousand Spectroscopically Selected Galaxy–Galaxy Gravitational Lens Candidates."

1. Problem Statement

Strong gravitational lensing is a critical tool for cosmology, allowing for the direct measurement of dark matter halo substructure, constraints on the Hubble constant ($H_0$) via time-delay cosmography, and the study of high-redshift galaxies. However, the number of confirmed strong lens systems remains low (approximately 400 confirmed systems) compared to the number of candidates (~$10^4$).

Existing discovery methods have limitations:

• Imaging Surveys: Effective for lenses with large Einstein radii ($\theta_E$) but miss systems where the lens and source are too close to be resolved by ground-based seeing.
• Spectroscopic Surveys: Previous spectroscopic searches (e.g., SDSS/SLACS, BELLS) were limited by fiber sizes (2"–3") and lower spectral resolution, often missing smaller systems or lacking the depth to find high-redshift background sources.
• Data Gap: There is a need for a large, spectroscopically confirmed catalog of lenses with precise redshifts for both the foreground lens and background source to facilitate future high-resolution follow-up and cosmological studies.

2. Methodology

The authors utilized data from the Dark Energy Spectroscopic Instrument (DESI) to perform a single-fiber spectroscopic lens search.

• Data Sample: The search was conducted on 5,837,154 luminous red galaxy (LRG) spectra from the DESI Year 3 internal data release (Loa), comprising the Main Survey and Survey Validation 3 (SV3). LRGs were selected because they are massive, quiescent, and have well-defined spectral energy distributions, making deviations (like background emission lines) easier to detect.
• Selection Strategy: The method relies on identifying background ionized oxygen [O II] $\lambda\lambda3727$ emission lines within the spectrum of a foreground LRG. Since DESI fibers are small (0.75" radius), the presence of a background emission line implies the background source is spatially coincident with the foreground galaxy, suggesting a lensing configuration.
• Three-Phase Pipeline:
  1. Residual Spectrum Generation: The best-fit galaxy model (generated by the REDROCK pipeline) is subtracted from the observed spectrum. Masks are applied to known foreground emission lines and sky lines to prevent false positives.
  2. Phase 1 (Linear Fit): A mock double-Gaussian profile (representing the [O II] doublet) is fitted to the residual spectrum across a grid of redshifts ($z_{BG}$) and velocity dispersions ($v_{disp}$). Candidates are selected if the fit significantly improves the $\chi^2$ (reduction > 100) and the background redshift is higher than the foreground redshift ($z_{BG} > z_{FG}$).
  3. Phase 2 (Non-linear MCMC): A secondary Markov Chain Monte Carlo (MCMC) fit refines the parameters (amplitudes, velocity dispersion) for the top candidates. Cuts are applied to ensure physical plausibility (e.g., positive amplitudes, line flux ratios consistent with star-forming galaxies, $20 < v_{disp} < 400$ km/s).
  4. Phase 3 (Contaminant Removal): Visual and statistical cuts remove "streak" contaminants caused by poorly fitted rare foreground emission lines (e.g., He II, [Fe VII]) and M-dwarf blends.
• Lensing Probability Calculation: For each candidate, the probability ($p_L$) that the system is a true lens rather than a chance alignment is calculated. This involves:
  • Modeling the foreground LRG as a Singular Isothermal Sphere (SIS).
  • Calculating the Einstein radius ($\theta_E$) using foreground/background redshifts and velocity dispersion.
  • Integrating the probability density function of the impact parameter over the lensing cross-section versus the annulus outside it, weighted by the [O II] luminosity function.

3. Key Contributions

• Catalog Creation: The paper presents a catalog of 4,110 strong galaxy–galaxy gravitational lens candidates.
• Novelty: 3,887 of these candidates are new discoveries, representing a ~50% increase in the total number of known strong lens candidates in the literature.
• Recovery of Known Systems: The search successfully recovered the host of the multiply imaged Type Ia supernova iPTF16geu, validating the method's ability to find small Einstein radius systems ($\theta_E \approx 0.3"$).
• Redshift Precision: Unlike imaging-based searches, this method provides precise spectroscopic redshifts for both the lens ($z_{FG}$) and the source ($z_{BG}$) for all candidates.
• Completeness Analysis: The authors quantified the pipeline's completeness, finding a maximum of ~70% for specific flux and width ranges, and noting that the catalog is likely incomplete for very faint sources ($<10 \times 10^{-17}$ erg s$^{-1}$ cm$^{-2}$).

4. Results

• Candidate Statistics:
  • Total Candidates: 4,110.
  • Redshift Range: Foreground lenses span $0.01 < z_{FG} < 1.40$; background sources span$0.33 < z_{BG} < 1.50$.
  • Einstein Radii: Calculated radii range from 0.1" to 4".
  • Probability: Based on the lensing probability calculation, the authors expect 53% (2,165) of the candidates to be true lenses. Specifically, 1,311 candidates have a probability $p_L > 0.9$.
• Comparison with Other Surveys:
  • Only 18 candidates overlap with the DESI Lens Foundry (imaging-based neural network searches), confirming that the single-fiber spectroscopic method selects a distinct population of lenses with smaller Einstein radii (often unresolved in imaging).
  • There is significant overlap with previous spectroscopic surveys (SLACS, BELLS, SILO), but the DESI sample extends to higher foreground redshifts ($z_{FG} > 1.0$) due to DESI's targeting strategy and fiber size.
• Imaging Confirmation:
  • The authors cross-matched candidates with the Euclid Quick Data Release (Q1). Two candidates showed obvious lensed arcs, confirming the method's validity.
  • Approximately 2,000 candidates fall within the footprint of the Vera C. Rubin Observatory (LSST), and ~1,300 within Euclid's Wide Survey, offering immediate paths for high-resolution confirmation.

5. Significance

• Dark Matter Substructure: This catalog provides the largest sample of galaxy-scale lenses with known redshifts, enabling statistical studies of dark matter substructure at cosmological distances via flux ratio anomalies.
• Cosmology ($H_0$): The sample is a prime target for monitoring multiply imaged transients (supernovae, quasars). Time-delay measurements from these systems will provide independent, high-precision constraints on the Hubble constant, helping to resolve the current $H_0$ tension.
• High-Redshift Astrophysics: The lensing magnification allows for the study of faint, star-forming galaxies at $z \sim 1.5$ that would otherwise be undetectable.
• Future Surveys: The catalog serves as a "target list" for upcoming space-based (Euclid, Roman) and ground-based (Rubin) surveys, maximizing the efficiency of follow-up observations for lens confirmation and transient monitoring.

In conclusion, this work significantly expands the known population of strong gravitational lenses, leveraging DESI's massive spectroscopic dataset and unique single-fiber selection technique to uncover a population of lenses that are inaccessible to traditional imaging surveys.