Search for Dark Photons between 16.96--19.52 $μ$eV with the HAYSTAC Experiment

Abstract

We report dark photon results from HAYSTAC Phase II using data from previously reported axion searches. Additionally, we present an analysis of an unpublished dataset covering a region between 19.46--19.52 $μ$eV. This region overlaps with a recently reported dark photon signal at 19.5 $μ$eV with a kinetic coupling strength of $|χ_{\text{rand}}| \simeq 6.5 \times 10^{-15}$ resulting from a reanalysis of previously published data from the TASEH collaboration. Given HAYSTAC's sensitivity, if such a signal were present, it would have appeared as a large $17.1σ$ excess above the noise. However, no such signal was observed. We thus exclude couplings $|χ_{\text{rand}}|\geq 4.90\times10^{-15}$ at the 90\% confidence level over the newly reported region. In addition, using our previously reported axion data, we exclude couplings $|χ_{\text{rand}}|\geq2.90\times 10^{-15}$ between 16.96--19.46 $μ$eV at the 90\% confidence level.

Figures (2)

• Figure 1: Top: Dark photon kinetic mixing for the fixed polarization scenario in the 1--45μ eV mass range, showing results from TASEH TASEH:2022TASEH_DP_2025, ADMX ADMX2024, CAPP CAPP8:2024CAPPMAX:2024, CAST-CAPP CASTCAPP:2022, ORGAN ORGAN:2024, QUAX QUAX:2024, SQuAD SQuAD:2020ymh, SQMS SQMS:2022gtv, Hefei Hefei:2024slu, SHANHE SHANHE:2023kxz (adapted from online repository cajohare:github). Also shown are HAYSTAC Phase I and Phase II, including exclusions converted by the method discussed in the text which uses $\langle \cos^2\theta \rangle=0.004$ (blue), and what is currently shown in cajohare:github (light green), which uses $\langle \cos^2\theta \rangle=0.019$ for a fair comparison with other results. Bottom: Exclusion on the dark photon kinetic mixing, comparing both the random (light fill) and fixed (solid fill) polarization scenarios, highlighting HAYSTAC Phase II (a)-(d) (blue), Phase II(e) (green) and TASEH (purple). The narrow vertical gaps mark the regions that are removed from the exclusion due to RFI contamination described in the text.
• Figure 2: Left: Zoomed in exclusion on the dark photon coupling with random polarization showing results from HAYSTAC Phase II(e) in green and TASEH results in purple. The baseline exclusion, shown in green, assumes the conservatively degraded form factor, and the yellow shaded region highlights the degradation of the exclusion from the 50% systematic uncertainty on the form factor. The red cross denotes the reported tentative dark photon signal from the TASEH reanalysis at 19.479727µ (4.710178) with a kinetic mixing strength of $|\chi_\text{rand}| \simeq 6.5 \times 10^{-15}$TASEH_DP_2025. The vertical grey line marks the injected signal discussed in the text. Right: Normalized power observed in the grand spectrum P$_{\text{GS}}$ (blue solid line) at each observed frequency $\nu$, shown in a zoomed window around the reported dark photon signal. The horizontal grey dashed line corresponds to HAYSTAC's 3.468$\sigma$ rescan threshold. The green dashed peak shows the expected signal strength of $17.1\sigma$, using the conservatively reduced form factor of 50%, had the tentative signal been present during HAYSTAC's data taking.