Bell-state measurement exceeding 50% success probability with linear optics

Abstract

Bell-state projections serve as a fundamental basis for most quantum communication and computing protocols today. However, with current Bell-state measurement schemes based on linear optics, only two of four Bell states can be identified, which means that the maximum success probability of this vital step cannot exceed $50\%$. Here, we experimentally demonstrate a scheme that amends the original measurement with additional modes in the form of ancillary photons, which leads to a more complex measurement pattern, and ultimately a higher success probability of $62.5\%$. Experimentally, we achieve a success probability of $(57.9 \pm 1.4)\%$, a significant improvement over the conventional scheme. With the possibility of extending the protocol to a larger number of ancillary photons, our work paves the way towards more efficient realisations of quantum technologies based on Bell-state measurements.