A new Limit for Axion Dark Matter with SPACE

TL;DR

This work reports a targeted axion dark matter search in Germany using the SPACE haloscope, scanning a narrow mass window around $m_a \approx 16.6\ \mu\mathrm{eV}$ with a 14 T magnet. Employing a high-quality TM$_{010}$ cavity in a resonant setup and a HAYSTAC-inspired analysis pipeline, the team found no axion signal and set 95% CL upper limits on the axion-photon coupling, achieving $g_{a\gamma\gamma}$ limits down to $2.811\times 10^{-13}\ \mathrm{GeV}^{-1}$ at peak sensitivity, two orders of magnitude better than prior constraints in this mass range. The experiment demonstrates rigorous calibration, noise characterization, and systematic error propagation, and situates the result as a substantial step toward the KSVZ benchmark (within a factor of ~44) while exceeding CAST limits. The methodology and sensitivity advance narrow-band axion searches and inform future cavity experiments operating at high magnetic fields.

Abstract

The axion, which has yet to be discovered, is a promising candidate for dark matter that emerges from Peccei-Quinn theory. This article presents the search for axion dark matter with the "Student Project for an Axion Cavity Experiment" (SPACE), which is also the first one in Germany. The hypothetical particle was looked for in the mass range from $16.626~\mathrm{μeV}$ to $16.653~\mathrm{μeV}$, corresponding to a frequency range of 4.020 GHz to 4.027 GHz, using a resonant cavity in a peak magnetic field of 14 T. No significant signal was found, allowing us to exclude an axion-photon coupling $g_{aγγ} = 14.6 \cdot 10^{-13}~\mathrm{GeV}^{-1}$ for the full mass range and $g_{aγγ} = 2.811 \cdot 10^{-13}~\mathrm{GeV}^{-1}$ at peak sensitivity with a 95% confidence level. This limit surpasses previous constraints by more than two orders of magnitude.

Figures (6)

• Figure 1: Cavity and receiver chain setup: Cylindrical resonant cavity in the magnet bore followed by a three-stage signal amplification receiver chain consisting of low noise amplifiers, bandpass filters and a mixing component for signal readout at the ADC.
• Figure 2: Reflection coefficient $\left|S_{11}\right|^2$ (dashed) and system noise temperature $T_\text{sys}$ (solid) of the setup as a function of frequency. The maximum signal power occurs at the minimum of $\left|S_{11}\right|^2$ and maximum of $T_\text{sys}$, respectively. It is slightly shifted between measurements because of changes in the physical environment. The residual $\Delta T_\text{sys}$ between the measured and fitted system temperature is shown in the lower plot.
• Figure 3: Magnetic field strength obtained from simulation inside and around the cavity (orange box) placed around $z=0$ such that the peak field strength of 14T is achieved inside the cavity.
• Figure 4: Change of physical temperature, loaded quality factor (running average), maximum system noise temperature (running average), and resonance frequency during the run.
• Figure 5: Normalized power excess (grand spectrum) as a function of frequency. No significant excess is detected.