Searching for an Oscillating Massive Scalar Field as a Dark Matter Candidate Using Atomic Hyperfine Frequency Comparisons

We use 6 yrs of accurate hyperfine frequency comparison data of the dual rubidium and caesium cold atom fountain FO2 at LNE-SYRTE to search for a massive scalar dark matter candidate. Such a scalar field can induce harmonic variations of the fine structure constant, of the mass of fermions, and of the quantum chromodynamic mass scale, which will directly impact the rubidium/caesium hyperfine transition frequency ratio. We find no signal consistent with a scalar dark matter candidate but provide improved constraints on the coupling of the putative scalar field to standard matter. Our limits are complementary to previous results that were only sensitive to the fine structure constant and improve them by more than an order of magnitude when only a coupling to electromagnetism is assumed.

Figure 1

Measurements of the normalized ratio of Rb/Cs frequencies at 864 s interval, spanning November 2009 to February 2016. Total of 100 814 points with mean $=1.1\times {10}^{-16}$ and standard deviation $=3\times {10}^{-15}$.

Figure 2

Top: Normalized power spectrum (blue) obtained from the best fit [see Eq. (10)] with the corresponding 5% detection threshold (see text). The red line corresponds to the maximum allowed signal at 95% confidence. Bottom: Non-normalized amplitude spectrum $\mathcal{A}$ (blue) and corresponding maximum allowed signal at 95% confidence (red).

Figure 3

Estimated values of the linear combination $d_e+k_q/k_{\alpha }(d_{\widehat{m}}-d_g)=d_e+0.043(d_{\widehat{m}}-d_g)$ of coupling constants $d_i$ between a massive scalar field and standard matter fields as a function of scalar field mass. The best fit values are shown in blue, with the 95% confidence upper bounds in red. The purple dashed line represents the 95% confidence upper bound obtained with Dy atoms in [2], which is only sensitive to $d_e$.