Achieving speedup in Dark Matter search experiments with a transmon-based NISQ algorithm

Abstract

Coherent detection of ultralight bosonic dark matter can be achieved by monitoring slow Rabi oscillations in superconducting qubits. We introduce an ancilla-assisted, gate-based protocol that enhances sensitivity to the hidden photon kinetic mixing parameter $ε$ using a single two-qubit gate, bypassing the need to maintain long-lived multi-qubit entangled states and remaining compatible with the limitations of modern quantum hardware. We characterized the increase in sensitivity accounting for decoherence, thermal occupation, errors in readout and reset, indicating up to a ten-fold reduction in the required integration time to reach the same exclusion limit on $ε$ achievable via Rabi-sampling experiments. Under plausible hardware assumptions and three years of data taking, the projected $95\%$ C.L. exclusion limit on the hidden photon mixing parameter reaches $ε\approx 1\times 10^{-14}$ across $2.5$-$6.0$ GHz ($10$-$25$ \textmu eV).

Figures (4)

• Figure 1: Enhanced direct dark matter search outline. A sensing qubit is prepared in $\ket{0_\text{S}}$ and exposed to the dark matter-induced electric field for a time $\tau$. The sensing qubit state evolution depends on whether the qubit is resonant or off-resonance with the field. Applying the enhancement circuit (requiring an ancilla qubit A) further separates the on-resonance and off-resonance states, improving state discrimination.
• Figure 2: Excitation probability for the baseline experiment $P^\mathrm{obs}$, the conditioned enhanced signal probability $\tilde{P}^\mathrm{obs}$ (given protocol success), and the enhancement success probability $P_s^\mathrm{obs}$, plotted as a function of the ideal sensing probability $P_e$. Results correspond to the example noise configuration $\beta = 0.4$, $T_1=100$ µ s, $r=p=0.01$. Dashed lines show the corresponding ideal (noise-free) probabilities.
• Figure 3: Speedup factor $\mathcal{G}$ of the two-qubit enhanced protocol relative to two parallel single-qubit baseline experiments, computed for $m_X=4.5$ GHz. The speedups are evaluated on an $r$-$p$ grid with both errors varying from $0.1\%$ to $5.0\%$. (a) Medium coherence regime, with $Q=\pi\times 10^6$ for both sensing and ancilla. (b) High coherence scenario, with $Q=2\pi\times 10^6$ for both sensing and ancilla.
• Figure 4: Smooth envelope of the projected 95% C.L. exclusion limits on the hidden-photon kinetic-mixing parameter $\epsilon$ for the baseline and enhanced protocols, on top of existing haloscope limits and cosmological bounds in the $1$-$10$ GHz band AxionLimits.