🔬 physics

Searching for Dark Photons and ALPs using Gamma-rays from Nuclear Reactors

This paper presents a detector-independent framework for searching for dark photons and axion-like particles using gamma rays from nuclear reactors and derives projected sensitivities for the CONUS+ and CONUS100 germanium experiments at the Leibstadt reactor, highlighting their potential to probe previously unexplored parameter spaces for these light, weakly coupled particles.

Original authors: Lucia Angel, Manfred Lindner, Farinaldo S. Queiroz, Bilge S. Solak

Published 2026-10-30

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Lucia Angel, Manfred Lindner, Farinaldo S. Queiroz, Bilge S. Solak

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

The universe is filled with matter that we cannot see. Astronomers know this because galaxies spin too fast to be held together by the visible stars and gas alone, and because light from distant objects bends around invisible clusters of mass. This missing substance, known as dark matter, makes up most of the matter in the cosmos, yet its true nature remains one of the greatest mysteries in physics. For decades, scientists focused their search on heavy, slow-moving particles, but as those searches have come up empty, attention has shifted toward lighter, more elusive candidates. Among the most compelling possibilities are "dark photons" and "axion-like particles." These are hypothetical, very light particles that might interact with ordinary matter only through extremely weak forces, acting as a hidden bridge between the visible world and the dark sector. Because they interact so faintly, finding them requires looking in places where the background noise is low and the source of potential signals is incredibly intense.

Nuclear reactors, which power cities by splitting atoms, provide such a source. Inside a reactor core, the splitting of atoms and the subsequent settling of nuclear fragments release a torrent of high-energy light particles called gamma rays. While these gamma rays are well understood, a new study suggests they might also be the birthplace of the elusive dark particles scientists are hunting. A team of researchers has now modeled how these gamma rays could transform into dark photons or axion-like particles and how a nearby detector could catch them. They focused their calculations on a specific experiment called CONUS+, which sits just twenty meters from a massive reactor in Switzerland, and on a proposed future upgrade to that experiment. Their work does not claim to have found these particles yet; rather, it maps out exactly where and how to look for them, defining the limits of what current and future equipment can detect.

The researchers began by treating the reactor as a factory for these hidden particles. In the case of dark photons, which are cousins to the ordinary photon of light but carry a tiny amount of mass, the process is akin to a game of billiards. A gamma ray from the reactor strikes an electron in the reactor's fuel, and instead of bouncing off, it transforms into a dark photon. This new particle then flies out of the reactor core, passing through the shielding and the surrounding rock with almost no interaction, until it reaches the detector. Inside the detector, which is filled with ultra-pure germanium crystals, the dark photon strikes another electron and transforms back into a gamma ray, depositing a specific amount of energy that the machine can record. The team calculated that for particles with masses below half a million electron volts, the current CONUS+ setup could detect a signal if the interaction strength is roughly one part in one hundred thousand. A larger, upgraded version of the experiment, called CONUS100, could push this sensitivity even further, reaching down to one part in eighty thousand.

The search for axion-like particles is more complex because these particles can be created in several different ways. Some are born when a gamma ray interacts with the electric field of an atomic nucleus, a process known as Primakoff conversion. Others are created when a gamma ray hits an electron, or when an excited atomic nucleus, instead of releasing a gamma ray, releases an axion-like particle. Once created, these particles travel to the detector, where they might scatter off an electron, be absorbed by an atom, or even decay into a pair of electrons and positrons if they are heavy enough. The researchers found that the current CONUS+ experiment is limited by its ability to read out energy signals only up to 0.6 million electron volts. This cutoff means it cannot see the heaviest axion-like particles or those that decay into electron pairs, which require higher energy thresholds to be spotted.

However, the proposed CONUS100 upgrade, which would increase the detector mass to one hundred kilograms and extend its energy reading capability up to ten million electron volts, opens a window into completely unexplored territory. The study shows that this larger detector could probe a region of particle masses around the million electron volt scale that has never been tested before. Specifically, it could look for axion-like particles that are produced in a very specific nuclear reaction involving hydrogen and neutrons, a reaction that creates a distinct, sharp line of energy at 2.23 million electron volts. If these particles exist and have the right properties, the upgraded detector could see them decaying into electron-positron pairs, a signature that would be impossible to miss.

The team also compared their projected sensitivity against existing limits from other experiments, such as those using particle accelerators or searching for solar axions. They found that for dark photons, the CONUS+ experiment alone could improve upon the best previous limits from a reactor experiment by a factor of roughly one and a half. For the axion-like particles, the situation is more nuanced. In the low-mass range where the particles live a long time, current experiments have already ruled out the simplest versions of the theory. But the study highlights a specific gap: a region where the particles are heavy enough to decay into electron pairs but light enough to survive the short trip from the reactor to the detector. In this narrow window, the CONUS100 experiment could be the first to find these particles, provided they exist.

The researchers emphasize that these results are projections based on detailed simulations of how the detectors would behave and how the particles would interact. They have not yet observed a signal; instead, they have drawn a precise map of where the signal would appear if it exists. The study relies on a simplified assumption of background noise to show the best-case scenario for what the detectors could achieve. In reality, a future analysis would need to account for the specific, measured background of the detector and the subtle variations in the reactor's operation. Nevertheless, the work demonstrates that nuclear reactors, long used to study neutrinos, are also powerful laboratories for hunting the lightest and most elusive members of the dark matter family. By pushing the boundaries of what these detectors can see, the team has shown that the next generation of reactor experiments could finally peer into the dark sector in ways that were previously thought impossible.

Technical Summary: Searching for Dark Photons and ALPs using Gamma-rays from Nuclear Reactors

Problem and Motivation
The microscopic nature of dark matter remains unknown, motivating searches for light, weakly coupled particles in the sub-GeV and MeV mass ranges. While Weakly Interacting Massive Particles (WIMPs) have been the traditional benchmark, the absence of conclusive signals has shifted focus toward dark sectors containing new mediators, such as dark photons and axion-like particles (ALPs). These particles pose distinct experimental challenges: their small couplings imply suppressed production and interaction rates, and their low masses often fall below the thresholds of conventional high-energy searches. Nuclear reactors offer a unique environment for such searches, providing intense fluxes of MeV gamma rays generated through fission, radiative neutron capture, and the de-excitation of fission products. These photons, along with specific magnetic nuclear transitions, can act as sources for light dark-sector particles, offering complementary probes to standard dark matter searches.

Methodology
This work presents a detector-independent framework for describing the production and detection of light dark-sector particles by reactor gamma rays, applied specifically to the CONUS+ experiment and a proposed 100-kg upgrade, CONUS100.

Source Modeling:
- Continuous Spectrum: The continuous reactor photon spectrum is modeled using an empirical parameterization up to 10 MeV, normalized to the reactor thermal power.
- Discrete Transitions: Six specific magnetic nuclear transitions are considered, including the 2.23 MeV line from proton-neutron capture ( $p(n, \gamma)d$ ) and four M4 transitions from fission fragments. Yields are taken from TEXONO data and scaled to the Leibstadt (KKL) reactor fission rate.
Dark Photon Scenario:
- Model: A kinetically mixed dark photon ( $A'$ ) with mass $m_{A'}$ and mixing parameter $\epsilon$ .
- Production: Reactor photons produce dark photons via Compton-like scattering on electrons ( $\gamma + e^- \to A' + e^-$ ). The calculation uses the exact tree-level cross-section and the total photon cross-section of uranium (including photoelectric absorption) rather than just the Compton fraction.
- Detection: Dark photons are detected via the inverse process ( $A' + e^- \to \gamma + e^-$ ) in a germanium detector. The analysis assumes $m_{A'} < 2m_e$ to forbid tree-level decays to $e^+e^-$ .
Axion-Like Particle (ALP) Scenario:
- Model: A generic ALP ( $a$ ) coupled to photons ( $g_{a\gamma\gamma}$ ), electrons ( $g_{aee}$ ), and nucleons ( $g_{ann}^{(0,1)}$ ).
- Production Mechanisms:
  - Photon Coupling: Primakoff conversion ( $\gamma + N \to a + N$ ) using the continuous spectrum.
  - Electron Coupling: Compton-like scattering ( $\gamma + e^- \to a + e^-$ ) using the continuous spectrum.
  - Nucleon Coupling: Emission from discrete magnetic nuclear transitions ( $N^* \to N + a$ ), competing with photon emission.
- Detection Mechanisms:
  - Photon Coupling: Inverse Primakoff scattering ( $a + N \to \gamma + N$ ) and $a \to \gamma\gamma$ decays within the detector.
  - Electron Coupling: Inverse Compton-like scattering, axio-electric absorption, and $a \to e^+e^-$ decays (for $m_a > 2m_e$ ).
  - Nucleon Coupling: Since germanium lacks the specific resonant isotopes for direct nuclear absorption of reactor ALP lines, detection relies on a second coupling (photon or electron) or in-detector decays.
Experimental Application:
- CONUS+: A 3.74 kg array of High Purity Germanium (HPGe) detectors at 20.7 m from the 3.6 GW KKL reactor, with a data acquisition (DAQ) range of 25 keV to 0.6 MeV.
- CONUS100: A hypothetical 100 kg upgrade with an extended DAQ range up to 10 MeV.
- Sensitivity Calculation: Projected sensitivities are derived assuming a flat background of 10 DRU (Dark Matter Units). Dark photon limits are quoted at 95% C.L., while ALP limits are at 90% C.L.

Key Results

Dark Photons:
- For masses $0.5 \text{ keV} \lesssim m_{A'} \lesssim 0.5 \text{ MeV}$ , CONUS+ reaches a kinetic mixing sensitivity of $\epsilon \simeq 1.65 \times 10^{-5}$ , and CONUS100 reaches $\epsilon \simeq 1.2 \times 10^{-5}$ .
- These projections improve upon the re-analyzed TEXONO reactor limit by factors of approximately 1.6 (CONUS+) and 2.2 (CONUS100).
- CONUS+ sensitivity deteriorates rapidly as $m_{A'}$ approaches its 0.6 MeV energy window limit, whereas CONUS100 maintains sensitivity up to the $A' \to e^+e^-$ threshold ( $2m_e$ ).
Axion-Like Particles (Single Coupling):
- Photon Coupling ( $g_{a\gamma\gamma}$ ): In the long-lived regime, sensitivity is mass-independent ( $|g_{a\gamma\gamma}| \simeq 9.6 \times 10^{-4} \text{ GeV}^{-1}$ for CONUS+). For $m_a \gtrsim 50 \text{ keV}$ , the $a \to \gamma\gamma$ decay channel dominates, improving sensitivity to $\sim 1.5 \times 10^{-5} \text{ GeV}^{-1}$ at 2 MeV for CONUS100.
- Electron Coupling ( $g_{aee}$ ): Similar mass-independent behavior is observed for scattering channels ( $|g_{aee}| \simeq 1.2 \times 10^{-5}$ for CONUS+). For $m_a > 2m_e$ , the $a \to e^+e^-$ decay channel opens, creating a distinct sensitivity region for CONUS100 down to $|g_{aee}| \simeq 9 \times 10^{-8}$ .
- Comparison to Limits: The single-coupling projections for CONUS+ and CONUS100 generally lie within regions already excluded by current laboratory and solar searches (e.g., NEON, BaBar, beam dumps).
Axion-Like Particles (Nuclear Production):
- Using the $p(n, \gamma)d$ line, the paper projects sensitivities to coupling products $|g_{a\gamma\gamma}g_{ann}^{(1)}|$ and $|g_{aee}g_{ann}^{(1)}|$ .
- Unexplored Regions: The analysis identifies two specific regions not covered by existing experimental limits:
  1. The $a \to e^+e^-$ decay window for $|g_{aee}g_{ann}^{(1)}|$ between $2m_e$ and 2.23 MeV, where CONUS100 reaches $|g_{aee}g_{ann}^{(1)}| \simeq 1.3 \times 10^{-15}$ .
  2. Regions in the $(g_{a\gamma\gamma}, g_{ann}^{(1)})$ and $(g_{aee}, g_{ann}^{(1)})$ planes at MeV masses where ALPs decay between the Sun and Earth (evading solar constraints) but survive the 20 m reactor baseline.

Significance and Claims
The paper claims that nuclear reactor experiments, typically optimized for neutrino physics, provide a powerful and complementary platform for searching for light dark-sector particles. The primary significance lies in the ability to probe production and detection mechanisms distinct from those in direct dark matter detection or collider searches.

The authors emphasize that while single-coupling ALP searches largely overlap with existing excluded regions, the combination of nuclear production (via magnetic transitions) and detection via secondary couplings allows access to unexplored parameter space, particularly for MeV-mass ALPs. The CONUS100 upgrade is highlighted as a critical step, as its extended energy window (up to 10 MeV) and increased mass are necessary to access the $a \to e^+e^-$ decay channel and the specific regions where ALPs decay between the Sun and Earth but survive the reactor baseline.

The authors maintain a modest tone regarding their results, noting that the presented sensitivities are "idealized projections" based on a flat background benchmark and unit efficiency. They acknowledge that future experimental analyses must incorporate measured energy-dependent backgrounds, detailed detector response, and systematic uncertainties to convert these physics reach estimates into definitive exclusion limits. Nevertheless, the work demonstrates that reactor-based experiments can significantly improve upon existing dark photon limits and open new avenues for ALP searches at MeV scales.

More like this