Sensitivity of a closed dielectric haloscope to axion dark matter

This paper presents a computationally efficient model for determining the sensitivity of closed dielectric haloscopes to axion dark matter, which is validated using MADMAX prototype data to establish the foundation for future searches in the 26–500 $\mu$eV mass range.

The Gist

The Great Axion Hunt: Tuning a Cosmic Radio

Imagine the universe is filled with a ghostly, invisible fog called Dark Matter. For decades, scientists have been trying to catch a glimpse of it. One of the leading suspects is a tiny, elusive particle called the axion.

The problem? Axions are shy. They rarely interact with anything. But, according to theory, if you put an axion in a strong magnetic field, it might "trip" and turn into a tiny flash of light (a photon). The catch? This light is incredibly faint and has a very specific, high-pitched frequency, like a note on a piano that is too high for human ears to hear.

This paper describes how the MADMAX team built a super-sensitive "radio" to catch these faint notes, and more importantly, how they figured out exactly how loud their radio is so they know if they missed a signal or if the axions just aren't there.

Here is the breakdown of their work, using some everyday analogies.

---

1. The Machine: A "Cosmic Echo Chamber"

To catch these axions, the team built a device called a dielectric haloscope. Think of it not as a solid metal box (like a traditional radio), but as a stack of pancakes (dielectric disks) sitting on a frying pan (a metal mirror), all inside a giant magnet.

• The Setup: When the axion fog passes through the magnet, it tries to turn into light. The stack of pancakes is designed to catch that light and bounce it back and forth, making it louder and louder through constructive interference.
• The Goal: It's like stacking mirrors in a hallway to make a whisper echo until it sounds like a shout. The team wants to make that "shout" loud enough for their receiver to hear.

2. The Problem: The "Black Box" Simulation

The machine is huge (relative to the tiny wavelength of the axion light). If you want to know exactly how loud the machine will be, you have to simulate the physics inside it on a computer.

• The Analogy: Imagine trying to simulate the airflow inside a hurricane using a spreadsheet. It's possible, but it takes a supercomputer and days of running time.
• The Issue: The MADMAX machine is so big and complex that running these "full physics" simulations for every tiny adjustment (like moving a mirror by the width of a human hair) is too slow and expensive. They needed a shortcut.

3. The Solution: The "Transmission Line" Shortcut

Instead of simulating the whole hurricane, the authors built a simple model based on transmission lines (the same math used to design coaxial cables for your TV).

• The Analogy: Instead of simulating every drop of water in a river, you just measure the speed of the current at a few key points and use a simple formula to predict the flow.
• How it works: They treated the stack of pancakes and the mirror like a series of pipes and valves. By measuring how the machine "reflects" a signal sent into it (like shouting into a cave and listening for the echo), they could tune their simple model to match reality.
• The Magic: This simple model is fast. It can run in seconds on a laptop, whereas the full simulation takes days.

4. Dealing with Imperfections: The "Wobbly Mirror"

In the real world, nothing is perfect. The mirror might be slightly tilted, or the pancakes might be a tiny bit warped. In a complex simulation, you'd have to model every scratch and dent.

• The Paper's Trick: The authors realized they didn't need to model the dents. Instead, they treated the dents as if they were just changing the "effective" properties of the machine.
• The Analogy: Imagine you are trying to tune a guitar, but the strings are slightly rusty. Instead of modeling the rust, you just turn the tuning pegs a little bit more to compensate. The model "absorbs" the imperfections by slightly adjusting the numbers for the distance between the pancakes.
• Result: They proved that even with a wobbly mirror or a warped disk, their simple model could still predict exactly how loud the signal would be, as long as they let the numbers "wiggle" a little to fit the data.

5. The Noise: Distinguishing the Signal from the Static

The real challenge isn't just hearing the signal; it's knowing if what you hear is an axion or just the machine's own static (noise).

• The Analogy: Imagine trying to hear a friend whisper in a crowded, noisy bar. You need to know exactly how loud the bar is (the background noise) to know if your friend is actually speaking.
• The Method: The team built a second model for the "receiver" (the part that listens). They measured how the receiver behaves when connected to different "standards" (like a perfect silence, a perfect echo, and a perfect match).
• The Combination: They glued the "Machine Model" and the "Receiver Model" together. This allowed them to predict exactly what the system temperature (the noise level) should look like. When they compared this prediction to the actual data, they could see if the machine was behaving normally or if something had shifted (like the mirror moving slightly due to temperature changes).

6. The Result: A Blueprint for the Future

By using this simple, fast model, the team successfully analyzed data from their prototype experiment (CB200) at CERN.

• What they found: They confirmed that their machine works exactly as predicted. They could calculate the "Boost Factor" (how much the machine amplifies the axion signal) with high precision, even without running expensive supercomputer simulations.
• Why it matters: This proves that they can build much larger versions of this machine in the future. If they have to simulate every inch of a giant machine, it would be impossible. But with this "shortcut" model, they can design massive, super-sensitive axion detectors that could finally catch the ghostly axion.

Summary

This paper is essentially a user manual for a complex physics experiment. It says: "We built a giant, complicated machine to catch invisible particles. Instead of using a supercomputer to understand how it works, we built a simple, fast math model that acts like a 'tuning guide.' We proved this guide is accurate even when the machine is slightly imperfect. This means we can now build even bigger, better machines to find Dark Matter."

Technical Summary

Here is a detailed technical summary of the paper "Sensitivity of a closed dielectric haloscope to axion dark matter."

1. Problem Statement

The search for axion dark matter (DM) in the mass range of $\sim 26$ to $500$$\mu$eV (corresponding to frequencies of$\sim 6$–$120$ GHz) faces a significant challenge: the signal power of a standard cavity haloscope scales inversely with the square of the frequency, limiting sensitivity at higher masses.

• Dielectric Haloscopes: The MADMAX experiment proposes using a stack of dielectric disks and a mirror to enhance axion-photon conversion via constructive interference, effectively decoupling the conversion volume from the photon wavelength.
• The Computational Bottleneck: While the conversion volume can be scaled up (large diameter) to increase sensitivity, simulating these large-scale 3D systems using full-wave Finite Element Methods (FEM) is computationally prohibitive ($> O(10^3 \lambda^3)$ domain size).
• The Calibration Challenge: Determining the "boost factor" ($\beta^2$)—the enhancement of signal power relative to a simple magnetized mirror—requires precise knowledge of the system's electromagnetic response. Direct measurement of the internal electric field is difficult and time-consuming, especially for "closed" boosters (encased in a metal cylinder) where boundary conditions restrict mode propagation.

2. Methodology

The authors present a simplified, computationally efficient model to determine the sensitivity of a closed dielectric haloscope (specifically the CB200 prototype). The methodology involves three main components:

A. Booster Model (Transmission Line Approach)

Instead of full 3D simulations, the booster is modeled as a 1D transmission line (TL) network using circuit simulation software (Advanced Design System - ADS).

• Structure: The model consists of alternating TL segments representing air gaps and sapphire disks, terminated by a mirror with finite conductivity.
• Parameters: The model fits physical dimensions (disk spacing $d_i$, taper length $L$) and effective material properties (refractive index $n$, dielectric loss $\tan \delta$) to measured reflectivity ($\Gamma$) data obtained via a Vector Network Analyzer (VNA).
• Boost Factor Calculation: Once the model reproduces the reflectivity, the boost factor is calculated by simulating the axion-induced current excitation within the TL network. A geometric overlap factor ($|\eta_A|^2 \approx 0.84$) is applied to account for the transverse field shape mismatch between the ideal TE$_{11}$ mode and the uniform axion current.

B. Receiver Noise Model

To account for the receiver's impact on the system temperature ($T_{sys}$), a noise model is developed:

• LNA Modeling: The Low Noise Amplifier (LNA) is modeled using correlated voltage ($V_n$) and current ($I_n$) noise sources.
• Calibration: Parameters are extracted by fitting the model to system temperature spectra measured with known standards (open, short, and matched loads).
• Integration: The booster and receiver models are connected via a transmission line of length $L_{con}$ (electrical distance). This combined model reproduces the standing wave patterns (oscillations) observed in the system temperature spectra.

C. Model Verification and 3D Effects

The 1D model is validated against full-wave 3D FEM simulations (COMSOL) to ensure it captures realistic imperfections:

• Mirror Tilt & Disk Deformation: The model demonstrates that 3D effects (e.g., mirror tilt up to $0.1^\circ$, disk planarity deviations of$\sim 50$$\mu$m) can be effectively absorbed by adjusting the "effective" parameters of the 1D model (specifically increasing the effective dielectric loss$\tan \delta$ and slightly shifting disk positions).
• Mode Identification: The paper details a strategy to isolate the fundamental TE$_{11}$ "booster mode" from higher-order modes (HOMs) by carefully tuning the distance between the taper and the first disk ($d_0$), ensuring a frequency separation of 50–100 MHz. Field mapping using a dielectric bead confirms the TE$_{11}$ nature of the resonance.

3. Key Contributions

1. Efficient Sensitivity Determination: The paper establishes a method to determine the boost factor and system sensitivity using minimal computational resources (circuit simulation) rather than expensive 3D FEM simulations.
2. Handling Imperfections: It demonstrates that realistic geometric imperfections (tilts, non-planarity) in a closed booster can be accurately modeled by fitting effective parameters, making the model robust for experimental data analysis.
3. Closed Booster Analysis: Unlike previous "open" booster analyses that relied on intrusive field probes, this work provides a complete non-invasive characterization method for a closed, cylinder-encased system using only port measurements (reflectivity and system temperature).
4. Uncertainty Quantification: A rigorous Monte Carlo propagation of uncertainties is performed, accounting for power calibration errors, fit uncertainties, field shape factors, and time-dependent mechanical instabilities (mirror drift).

4. Results

The methodology was applied to data from the CB200 prototype (three 200 mm sapphire disks in an aluminum cylinder) operating in the Morpurgo magnet at CERN (1.6 T field).

• Data Runs: Five physics runs were analyzed across two configurations (centered at $\sim 18.55$ GHz and $\sim 19.21$ GHz).
• Model Accuracy: The model successfully reproduced the measured reflectivity ($|\Gamma|^2$) and group delay ($\tau_{gd}$) within uncertainties. It also accurately reproduced the system temperature spectra, including the broadband standing wave oscillations and the narrowband booster mode resonance.
• Boost Factor: The setup achieved boost factors of up to $\beta^2 \approx 2500$.
• Uncertainty: The combined relative uncertainty on the boost factor was determined to be between 9% and 14%, depending on frequency.
• Stability: The analysis identified mechanical instabilities in the mirror tuning mechanism, causing frequency drifts of $\sim 1$ MHz over time, which were corrected by adjusting the effective mirror-disk distance ($d_3$) in the model.

5. Significance

• Foundation for MADMAX: This work underpins the first axion dark matter search using a dielectric haloscope (published in Phys. Rev. Lett. 135, 041001). It provides the rigorous statistical framework required to set exclusion limits on the axion-photon coupling ($g_{a\gamma}$).
• Scalability: By proving that sensitivity can be calculated without full 3D simulations, the method enables the design and analysis of much larger, more complex boosters (with more disks and larger diameters) that would be computationally intractable to simulate otherwise.
• Operational Efficiency: The ability to relate system changes (e.g., mirror drift) directly to model parameters allows for real-time adjustments and data validation without interrupting data taking, a critical feature for long-duration dark matter searches.