⚛️ phenomenology

Probing multi-state dark matter via optical absorption lines in DESI spectra

By stacking 145,087 DESI spectra paired with foreground galaxy groups, researchers found no evidence for multi-state dark matter absorption lines, thereby establishing stringent upper limits on magnetic dipole transition cross-sections for MeV-scale dark matter in a parameter space inaccessible to current cosmological and direct detection experiments.

Original authors: Anoma Ganguly, Shadab Alam

Published 2026-09-16

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Anoma Ganguly, Shadab Alam

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

For decades, the most dominant substance in the universe has remained invisible. We know dark matter exists because its gravity holds galaxies together and shapes the cosmic web, yet every attempt to catch a particle of it has come up empty. Scientists have spent years searching for dark matter that might bump into ordinary atoms or scatter off electrons, but so far, no confirmed signal has emerged. This silence has led researchers to consider a different possibility: perhaps dark matter is not a single, static type of particle, but a family of particles that can change their state. Just as an atom can jump between energy levels, these dark matter particles might have a ground state and a slightly heavier excited state. If this were true, and if the particles could interact with light through a subtle magnetic connection, they might leave a faint, specific mark on the light passing through them.

A team of researchers has now tested this idea by looking for a ghostly shadow cast by dark matter on the light of distant galaxies. They focused on a scenario where a dark matter particle absorbs a photon of a very specific color to jump from its ground state to an excited state. If such particles exist in the halos of galaxy groups, they should act like a filter, removing a tiny slice of light from the background sources behind them. The challenge is that this effect is incredibly weak; a single galaxy group is far too faint to reveal the signal. To solve this, the researchers turned to a massive spectroscopic survey called DESI, which captures the light of millions of galaxies, and paired it with a catalog of nearby galaxy groups from the GAMA survey. By aligning the light from over 145,000 background sources with the positions of foreground galaxy groups, they effectively stacked the data, allowing a universal signal to emerge from the noise while random errors canceled each other out.

The result of this massive stacking effort was a clean, empty spectrum. When the researchers looked for the specific absorption line that would indicate dark matter jumping to an excited state, they found nothing. The data showed no dip in the light, meaning the optical depth—the measure of how much light was blocked—was consistent with zero. While this might sound like a null result, it is a powerful discovery because it allows the team to set strict limits on how strongly dark matter can interact with light. They determined that if these multi-state dark matter particles exist with masses in the range of a few million electronvolts and transition energies between 2.5 and 5 electronvolts, their ability to absorb light must be far weaker than previously untested theories allowed. This range is particularly important because it sits in a gap where other methods, such as studying the early universe or using direct detection experiments, cannot easily look.

The study confirms that the spectroscopic stacking method works as a viable tool for hunting these elusive particles. Although the current data from the early release of the DESI survey did not find the dark matter signal, the precision of the measurement is already strong enough to rule out certain types of interactions that were previously unknown. The researchers project that as the full DESI survey and future generations of telescopes collect more data, the sensitivity of this technique will improve dramatically. With more spectra to stack, the limits on dark matter's magnetic interactions could tighten by more than ten times, potentially closing the door on this specific type of dark matter or, if the signal is just below the current threshold, finally revealing its presence. For now, the universe remains silent in this specific frequency, but the method has proven that we can listen with a clarity that was previously impossible.

Technical Summary: Probing Multi-State Dark Matter via Optical Absorption Lines in DESI Spectra

Problem and Motivation
The particle nature of dark matter (DM) remains unresolved, with decades of searches for non-gravitational interactions yielding no confirmed signals. While standard DM is often modeled as a single-state particle, multi-state models offer a viable alternative where DM consists of multiple states that are electrically neutral but couple to photons via higher-order electromagnetic moments. Specifically, this paper investigates a minimal two-state system comprising a ground state ( $\chi$ ) and a nearly degenerate excited state ( $\chi^*$ ), connected by a transition magnetic dipole moment. In this framework, the radiative transition $\chi + \gamma \to \chi^*$ allows ground-state DM to absorb photons of energy $\Delta E$ . For transition energies in the range $\Delta E \sim \mathcal{O}(\text{eV})$ , this absorption manifests as a narrow, monochromatic line in the optical band. The authors aim to detect this signature by searching for absorption imprints on the spectra of background sources whose lines of sight pass through foreground dark matter halos.

Methodology
The study employs a stacking analysis to overcome the weakness of the signal from any single halo, which is undetectable in individual spectra. The methodology proceeds in three stages:

Data Construction: The authors construct source–absorber pairs using two datasets:
- Absorbers: Galaxy groups from the Galaxy And Mass Assembly (GAMA) survey (specifically regions G02, G09, G12, and G15). Groups are selected with a friends-of-friends multiplicity $N_{\text{fof}} > 3$ to ensure reliable dynamical mass estimates.
- Sources: Background galaxies and quasars from the Dark Energy Spectroscopic Instrument (DESI) Early Data Release (EDR).
- Pair Selection: Pairs are retained if the angular separation between the source and the absorber center is within the radius $R_{100}$ (enclosing 100 times the critical density) and if the source redshift exceeds the absorber redshift by $\Delta z > 0.02$ . This redshift buffer ensures the source is not a member of the foreground group, preventing coherent contamination from atomic/molecular lines of group members.
Spectral Processing and Stacking:
- Source spectra are shifted to the rest frame of their respective foreground absorbers ( $\lambda_{\text{obs}} / (1+z_{\text{abs}})$ ).
- Spectra are interpolated onto a common grid and combined using inverse-variance weighting to suppress uncorrelated noise.
- A smooth continuum is estimated via Fourier transform and subtracted to isolate narrow features. The stacked optical depth is calculated as $\hat{\tau} = 1 - \bar{F}/\bar{F}_c$ .
- Noise is estimated empirically using the 16th and 84th percentiles of the optical depth distribution in overlapping wavelength bins.
Constraint Derivation: The observed stacked optical depth is compared against a theoretical prediction derived from the magnetic dipole transition cross-section ( $\sigma_M$ ). The theoretical optical depth scales as $\tau \propto M^2/m_\chi$ (where $M$ is the magnetic dipole moment and $m_\chi$ is the DM mass). Upper limits are placed by requiring the predicted signal to remain below the $2\sigma$ noise envelope of the data.

Key Contributions and Results

First Observational Limits: The analysis of 145,087 source–absorber pairs from DESI EDR and GAMA yields a stacked optical depth consistent with zero across the optical wavelength range. Consequently, the authors establish the first $2\sigma$ upper limits on the magnetic dipole transition cross-section ( $\sigma_M$ ) for multi-state DM in the optical regime.
Parameter Space Coverage: The constraints are strongest for transition energies $\Delta E \sim 3\text{--}5$ eV and dark matter masses in the few MeV range ( $m_\chi \sim \text{few MeV}$ ). This specific region is identified as inaccessible to current cosmological probes (Big Bang Nucleosynthesis and Cosmic Microwave Background anisotropies) and direct detection experiments (which face kinematic thresholds for up-scattering in this mass/energy regime).
Forecasting: By analyzing the scaling of pair counts with survey area and the signal-to-noise improvement with the number of stacked spectra ( $\propto \sqrt{N}$ ), the authors forecast that full DESI (Year 5) and next-generation surveys (e.g., Spec-S5) will tighten these limits by more than an order of magnitude.

Significance
The paper claims that spectroscopic stacking represents a competitive and distinct probe for electromagnetically interacting multi-state dark matter. Unlike emission-based searches, this absorption technique is sensitive to dark matter across a wide range of foreground halo masses and density profiles, as the signal depends linearly on column density rather than the densest emitting regions. The authors emphasize that this method targets a "blind spot" in current DM searches—the intermediate mass/energy regime where neither cosmological constraints nor direct detection experiments are effective. Furthermore, the technique is presented as model-agnostic regarding the specific interaction mechanism, applicable to any scenario predicting new spectral signatures, thereby offering a powerful complementary route to uncovering dark matter interactions as spectroscopic datasets grow.

More like this