An Improved RF Cavity Search for Halo Axions

TL;DR

This work advances the ADMX RF cavity approach to search for halo axions in the mass window $m_a\in$[$1.9,3.3]\ \mu\mathrm{eV}$ by employing a tunable, high-Q cavity within a $7.9\ \text{T}$ field and a low-noise cryogenic receiver. The team implements two parallel data analyses, including a Wiener-filter method that weights frequency bins by the expected signal-to-noise content, improving sensitivity over the previous 6-bin approach. Across extensive integration and environmental checks, no axion signal is observed, yielding 90% CL upper limits on the axion-photon coupling $g_{a\gamma\gamma}$ and on the local axion dark matter density for both KSVZ and DFSZ models. The results tighten constraints on the halo axion parameter space and demonstrate a significant speedup and sensitivity gain from hardware and analysis improvements, quantified as about a factor of 2 faster scans and ~13% additional sensitivity from the Wiener filter. The study thus narrows viable axion models in the explored mass range and informs future, more sensitive haloscope searches.

Abstract

The axion is a hypothetical elementary particle and cold dark matter candidate. In this RF cavity experiment, halo axions entering a resonant cavity immersed in a static magnetic field convert into microwave photons, with the resulting photons detected by a low-noise receiver. The ADMX Collaboration presents new limits on the axion-to-photon coupling and local axion dark matter halo mass density from a RF cavity axion search in the axion mass range 1.9-2.3 microeV, broadening the search range to 1.9-3.3 microeV. In addition, we report first results from an improved analysis technique.

Figures (7)

• Figure 1: Sketch of the RF cavity axion detector.
• Figure 2: Schematic diagram of the receiver chain.
• Figure 3: Receiver power sensitivity (log$_{10}$$\delta$P/P) versus number of averages (log$_{10}$ N) [lower abscissa] or averaging time (log$_{10}$ days) [upper abscissa].
• Figure 4: Wiener filter weighting for axion mass corresponding to 500 MHz. The normalization is discussed in the text.
• Figure 5: Search confidence for finding thermalized axions (solid), narrow 1-bin wide axions (dotted), and wide 10-bin wide axions (dashed), from WF and 6-bin searches.