⚛️ high-energy experiments

Comment on the preprint: First Limits on Axion Dark Matter from a DALI Prototype arXiv:2603.21951

This paper presents an independent assessment of the DALI prototype's sensitivity to axion-like particle dark matter, concluding that no electromagnetic modes below 12 GHz exhibit significant coupling to the axion field sufficient to explain the reported sensitivity, and offers recommendations for future dielectric haloscope experiments.

Original authors: Erika Garutti, Anton Ivanov, Béla Majorovits

Published 2026-07-31

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Erika Garutti, Anton Ivanov, Béla Majorovits

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 is a giant, invisible ocean, and we are standing on a beach trying to figure out what's swimming in it. For decades, scientists have been looking for a very specific, ghostly creature called "dark matter." We know it's there because it has gravity—it holds galaxies together like invisible glue—but we can't see it, touch it, or hear it. One leading theory suggests this dark matter might be made of tiny, wobbly particles called "axions." These axions are so shy that they rarely interact with anything, but if you shine a strong magnetic field on them, they might turn into a tiny, faint flash of light, like a firefly blinking in a dark room.

To catch these fireflies, physicists build special traps called "haloscopes." These aren't nets made of string, but rather stacks of super-smooth mirrors and plates designed to amplify that tiny flash of light into a signal we can hear. It's like tuning a radio to find a specific station; if the radio is tuned just right, the static clears, and you hear the music. The challenge is that we don't know exactly what "station" (or frequency) the axions are broadcasting on, so scientists have to scan through a huge range of frequencies, hoping to catch a whisper in the cosmic static.

Now, a group of researchers known as the MADMAX collaboration has been watching a different team, the DALI collaboration, who recently claimed to have built a prototype trap called DALI. The DALI team reported that their setup was so sensitive it could have caught these axion fireflies in a specific range of frequencies, between 5 and 7.5 gigahertz. This sounded like a huge step forward. However, the MADMAX team, who are experts in building these very same types of traps, decided to take a closer look. They didn't just take the DALI team's word for it; they built a virtual version of the DALI machine on their computers to see if the physics actually worked as described.

What the MADMAX team found was a bit of a reality check. When they simulated the DALI setup using the exact numbers and descriptions provided in the DALI report, the "radio" didn't seem to be tuned to the right station at all. Their computer models showed that while the DALI machine did have some resonant frequencies (like notes a guitar string can play), the specific way the electric fields moved inside the machine meant it was terrible at catching the axion signal. In fact, for every frequency they tested between 5 and 7.5 gigahertz, the "catching power" was so weak it was practically zero.

The team pointed out a few specific reasons why the signal was lost. First, the electric fields inside the machine were bouncing back and forth in a way that canceled themselves out, like two people pushing a swing in opposite directions at the exact same time—the swing just sits still. Second, the resonance the DALI team claimed to see at 6.907 gigahertz didn't match up with the simulation; the closest match in their model was around 6.58 gigahertz, and even that one had almost no ability to grab an axion. The MADMAX team also noted that the original report didn't show enough data to be sure, such as a full view of the frequency spectrum to prove there weren't other weird signals hiding nearby.

Ultimately, this paper doesn't say the DALI experiment is a failure, but it does say that based on the information available, the claimed sensitivity to axion dark matter cannot be reproduced. The MADMAX team concludes that they cannot confirm the DALI setup is ready to hunt for axions in the way they described. Instead of celebrating a discovery, they are offering a set of strict rules and recommendations for how to properly test these machines in the future. They suggest that before anyone claims to have found a new way to catch dark matter, they need to clearly show how their machine couples to the signal, measure the full range of frequencies to see the whole picture, and prove that the machine is actually tuned to the right note. Until those checks are done, the axion fireflies remain elusive, and the hunt continues.

Technical Summary: Comment on "First Limits on Axion Dark Matter from a DALI Prototype"

Problem Statement
This paper presents an independent assessment of the sensitivity claims made by the DALI collaboration in their preprint [1] regarding the detection of axion-like particle (ALP) dark matter. The central issue addressed is the validity of the reported sensitivity in the 5 to 7.5 GHz frequency range. The authors, E. Garutti, A. Ivanov, and B. Majorovits (representing the MADMAX collaboration), aim to reproduce the DALI results using established formalisms to determine if the described apparatus can effectively couple to the axion-induced field. They specifically investigate whether any electromagnetic modes below 12 GHz possess a significant coupling to the axion field sufficient to account for the sensitivity reported by DALI.

Methodology
The authors employed a numerical simulation approach based on the geometric and material parameters explicitly described in the DALI preprint.

Simulation Setup: Using COMSOL, the team modeled the resonator stack consisting of 20 dielectric plates with a fixed spacing of 6.21 mm, enclosed by a perfect lossless mirror and a perfect antenna (port boundary).
Idealized Assumptions: The simulation assumed an idealized setup with unbound infinite disks, perfectly plane-wave-like electromagnetic fields, no edge scattering, and no transverse higher-order mode conversion. The electric field was polarized tangentially to the plate surface.
Material Parameters: The plates were modeled with a relative dielectric constant ( $\epsilon_r$ ) of 30, a loss tangent ( $\tan\delta$ ) of $10^{-4}$ , and a thickness of 1 mm.
Analysis Metrics: The authors evaluated the reflection response ( $S_{11}$ ) and group delay (GD) to identify resonances. Crucially, they calculated the power coupling of these resonant modes to the axion-induced field using the normalized form factor ( $C$ ), defined by the overlap integral between the external homogeneous magnetic field ( $B_e$ ) and the standing wave electric-field eigenfunction ( $E_m$ ).

Key Results
The simulation identified eight distinct resonances within the 5 to 7.5 GHz range, with frequencies ranging from approximately 5.18 GHz to 6.58 GHz.

Form Factor Analysis: The calculated form factors ( $C$ ) for all identified resonances were found to be marginal. The values ranged from $1.67 \times 10^{-6}$ to $2.8 \times 10^{-4}$ , with the highest value being $0.0028$.
Field Distribution: Analysis of the electric field distributions (visualized in Figure 3 and Figure 4) revealed that for the resonance closest to the DALI-reported frequency (approx. 6.58 GHz vs. DALI's 6.907 GHz), the positive and negative lobes of the field practically cancel out. This results in negligible overlap with the constant external magnetic field, explaining the extremely low form factors.
Frequency Range: No resonances were found in the 7.5 to 12 GHz range, consistent with the periodic nature of the geometry.
Discrepancy: The authors noted a frequency shift between their simulation (closest resonance at ~6.58 GHz) and the DALI report (6.907 GHz), attributing this to common variations in dielectric constants and geometric tolerances, which prompted their broad frequency scan.

Conclusions and Claims
Based on the available information in the DALI preprint and using standard formalisms, the authors conclude that they cannot reproduce the sensitivity reported by DALI for ALP or hidden photon dark matter.

Primary Claim: There is no mode below 12 GHz with significant coupling to the axion field that could account for the reported sensitivity. The power coupling for the identified modes is negligible.
Critique of DALI Preprint: The authors state that the DALI preprint lacks sufficient information to fairly evaluate its results. Specifically, the calculation of the axion-field coupling is not transparently described, and uncertainties are not adequately addressed.

Significance and Recommendations
The paper serves as a critical peer review of a claimed experimental result in the field of axion dark matter searches. Its significance lies in highlighting the necessity for rigorous validation of dielectric haloscope performance. The authors provide specific recommendations for future assessments in this field:

Transparency in Coupling: The calculation of axion-field coupling to specific modes must be transparently described, with uncertainties included in sensitivity estimations.
Qualitative Confirmation: Expected field shapes of reflectivity-induced modes should be confirmed qualitatively, for instance, using the bead pull method.
Broadband Reflectivity: Reflectivity measurements should be displayed over a broad frequency range to reveal other potential features or modes, rather than focusing solely on a single peak.
Data Presentation: Initial calibrated power spectra should be displayed alongside normalized power excess data over extended frequency ranges.

Ultimately, the authors assert that without these clarifications and validations, the sensitivity claims of the DALI prototype cannot be accepted based on current published formalisms.

Technical Summary: Comment on "First Limits on Axion Dark Matter from a DALI Prototype"

More like this