Search for post-inflationary QCD axions with a quantum-limited tunable microwave receiver

The QUAX experiment utilized a quantum-limited tunable microwave receiver to scan a frequency range around 10.2 GHz, successfully ruling out viable hadronic axion models in the preferred post-inflationary mass region above 40 μeV by finding no signal candidates.

The Gist

Imagine the universe is filled with a mysterious, invisible substance called dark matter. For decades, scientists have suspected that a tiny, ghostly particle called the axion might be the main ingredient of this dark matter. The axion is like a "cosmic ghost" that was created right after the Big Bang and has been drifting through space ever since.

The problem is, these ghosts are incredibly hard to catch. They don't shine, they don't bounce off things, and they barely interact with normal matter. However, there is a tiny chance that if a cosmic axion bumps into a strong magnetic field, it might briefly turn into a tiny spark of microwave light (a photon).

The Experiment: A Cosmic Radio Tuner

The team behind this paper, called QUAX, built a giant, high-tech "radio tuner" to listen for these sparks.

1. The Trap (The Cavity): They built a hollow copper cylinder, about the size of a large trash can, and placed a sapphire crystal inside it. Think of this like a musical instrument (like a flute) that is designed to resonate at a very specific pitch. In this case, the "pitch" is a microwave frequency around 10.2 GHz.
2. The Magnet (The Spark Generator): They placed this cylinder inside a massive magnet that is 8 times stronger than a standard MRI machine. This strong magnetic field is the "spark generator" that gives the axions a chance to turn into light.
3. The Tuning Knob: The tricky part is that scientists don't know exactly what "pitch" (mass) the axion has. It could be slightly higher or lower. So, the QUAX team built a special mechanism to physically squeeze and stretch the copper cylinder, allowing them to "tune" the radio to different frequencies, scanning a range of about 38 MHz.

The Super-Sensitive Ear (The Receiver)

Listening for a ghost is hard because the signal is so faint it's almost non-existent. To solve this, QUAX used a quantum-limited receiver.

Imagine trying to hear a pin drop in a hurricane. Most microphones would just hear the wind. But QUAX used a special amplifier (called a TWPA) that is cooled down to near absolute zero (colder than outer space). This amplifier is so sensitive it can hear the "whisper" of a single particle of light without adding its own noise. It's like having an ear that is perfectly quiet, allowing it to detect the faintest cosmic signal.

The Hunt: What They Found

The team spent about 225 hours scanning a specific slice of the frequency spectrum (centered around 10.2 GHz). This corresponds to an axion mass that scientists think is very likely to exist based on recent computer simulations of the early universe (specifically, a "post-inflationary" scenario).

The Result: They didn't find the axion.

However, in science, a "no signal" result is still a huge discovery. It's like searching a specific room in a haunted house with a super-sensitive ghost detector and finding nothing. You can now say with 90% confidence: "If axions exist in this specific mass range, they are not as 'loud' (as strongly coupled to light) as our best theories predicted."

Why This Matters

Before this experiment, there was a "preferred zone" for axions (masses above 40 micro-electron-volts) where many scientists thought the ghost would be hiding. The QUAX team scanned this zone with a sensitivity that was good enough to catch the most popular types of axion models (known as KSVZ and DFSZ models).

Because they found nothing, they have effectively ruled out those specific models for that mass range. It's like narrowing down a suspect list: "We know the ghost isn't wearing a red hat in this room."

The Bottom Line

The QUAX experiment successfully built a quantum-supercharged radio tuner and scanned a specific, high-priority area of the universe for dark matter axions. They didn't find the axion, but they proved that if it is there, it's hiding in a way that is even more elusive than our current top theories suggested. This forces scientists to rethink their models or look in even more difficult places to find the missing piece of the dark matter puzzle.

Technical Summary

Technical Summary: Search for Post-Inflationary QCD Axions with a Quantum-Limited Tunable Microwave Receiver

Problem and Motivation
The axion was proposed to resolve the strong CP problem in quantum chromodynamics (QCD) via the Peccei-Quinn (PQ) symmetry [1–3]. It is a well-motivated dark matter candidate, yet its mass remains unknown, spanning several orders of magnitude [7]. While experimental searches have successfully probed axion-photon couplings in the mass region below 30 µeV (notably by ADMX and CAPP reaching DFSZ sensitivity, and HAYSTAC reaching near-KSVZ sensitivity), searches at higher masses have been hindered by unfavorable scaling of cavity haloscope parameters.

Recent lattice simulations [21, 22] motivate a specific search for post-inflationary axions with masses above 40 µeV. However, no extended searches have previously been conducted in this mass region at the sensitivity required to probe the KSVZ or DFSZ hadronic axion models. The QUAX (QUaerere AXion) collaboration aims to address this gap, specifically targeting the 35–45 µeV mass window (8.5–11 GHz).

Methodology and Experimental Apparatus
The experiment utilizes an improved high-frequency haloscope located at the Laboratori Nazionali di Legnaro (LNL). The apparatus consists of a tunable resonant cavity immersed in a strong magnetic field, coupled to a quantum-limited detection chain.

• Cavity Design: To enable wide frequency tuning at high frequencies, a new dielectrically loaded cavity was developed. It features a copper cylinder (414.3 mm length, 60.5 mm diameter) containing a sapphire cylinder (420 mm length, 30.2 mm outer radius). The science mode is the TM030. Frequency tuning is achieved via a "clamshell" mechanism where the copper cylinder is split into two halves; a computer-controlled linear positioner acts on a pantograph to open the halves up to one degree, tuning the frequency from 10.212 GHz down to 10.154 GHz. Finite element simulations and bead-pull measurements determined the cavity form factor $C_{030} \approx 0.40\text{--}0.43$.
• Cryogenics and Magnetic Field: The cavity operates inside a wet dilution refrigerator at approximately 100 mK to ensure thermal stability. The system is subjected to a vertical solenoidal magnetic field of 8.0 T (generated by a 92 A current), providing an integrated $B^2$ value of 50.2 T$^2$ m over the cavity length.
• Detection Chain: The readout employs a quantum-limited traveling wave parametric amplifier (TWPA) as the first stage, operating at ~120 mK. This is followed by a cryogenic HEMT at the 4 K stage and a room-temperature HEMT. The TWPA gain is optimized at each frequency step to minimize system noise temperature ($T_{sys}$). The output is down-converted, filtered, and sampled at 4.4 MS/s.
• Data Acquisition: Two run sessions (June and November 2024) yielded an integrated acquisition time of ~225 hours over 18 days, with a data collection efficiency of ~50%. The search window covered approximately 38 MHz (65% of the available tuning range), limited by intruder modes and mechanical constraints in the 10.175–10.196 GHz interval.

Data Analysis
Data processing involved dividing files into 4 ms chunks for Fast Fourier Transform, resulting in a bin width of ~268 Hz. Preprocessing included removing mixer-level interference and normalizing spectra based on gain stability checks and cavity resonance frequency fitting (using Fano-like functions).

To search for axion signals, the team estimated the baseline using a Savitzky-Golay filter and computed confidence belts via Monte Carlo simulations. The signal power was modeled based on the KSVZ coupling constant ($g_{a\gamma\gamma}^{KSVZ} \approx 1.67 \times 10^{-14} \text{ GeV}^{-1}$ at 10.2 GHz) and the local dark matter density ($\rho_a \sim 0.45 \text{ GeV/cm}^3$). A detection threshold of Signal-to-Noise Ratio (SNR) = 4.5 was set to achieve a false alarm rate of one per month.

Results
No signal candidates were observed in the scanned frequency region. Consequently, the experiment set 90% confidence level (C.L.) exclusion limits on the axion-photon coupling ($g_{a\gamma\gamma}$).

• Exclusion Regions: Hadronic axion models with anomaly coefficient ratios $E/N = 44/3$ and $E/N = 29/3$ are excluded in mass intervals of 153 neV and 136 neV, respectively, located between 41.991 µeV and 42.234 µeV.
• Sensitivity: The experiment achieved a typical sensitivity allowing the probe of axion-photon coupling values predicted within hadronic QCD axion models. For the longest single acquisitions, the sensitivity reached the level of the KSVZ model. The system noise temperature reached a minimum of ~1.1 K (2.3 photons), representing the lowest noise temperature obtained for a tunable haloscope operating above 40 µeV.

Significance and Outlook
This work represents a significant advancement in the search for dark matter axions in the well-motivated post-inflationary region ($m_a \gtrsim 40$ µeV). The authors claim that the successful operation of a dielectrically loaded tunable cavity with a large effective volume, combined with a quantum-limited, wide-bandwidth detection chain in a strong magnetic field, demonstrates the feasibility of scanning this frequency range with KSVZ-class sensitivity.

The paper modestly notes that while no signal was found, the apparatus has laid the foundation for a tunable haloscope capable of covering the proposed 9.5 to 11 GHz range. Future improvements are foreseen, including the use of a more powerful magnet and an increase in duty cycle to enhance sensitivity and scan rates.