Bell's inequality allows testing experimentally whether nature can be described in a local-realistic way by measuring correlations between two separated systems. While being conceptually simple, a conclusive test sets stringent experimental requirements in order to exclude all possible local-realistic descriptions (avoid loopholes). These requirements include a high detection efficiency in the correlation measurements, as well as their strict space-like separation. Thus, although experimental tests started already in the early 70s, fulfilling all requirements became possible only very recently.

I will present our approach based on heralded entanglement of single trapped Rb atoms over a large distance combined with highly reliable and fast readout of their internal spin state. Our results show a clear violation allowing to exclude all local-realistic theories with a very high level of significance. Reliable violations of Bell's inequality, beyond their fundamental importance, also form the basis for certified generation of random numbers and device-independent quantum key distribution.


Seminar, September 16, 2016, 12:00. ICFO's Seminar Room

Hosted by Prof. Morgan Mitchell