Event–Ready Entanglement Preparation

All Bell experiments carried out so far have had ten or more times fewer coincidence counts than singles counts and this, in effect, means a detection efficiency under 10%. Therefore, all these experiments relied only on coincidence counts and herewith on additional assumptions. Recently, however, Santos devised hidden variable models which do not obey the assumptions and thus made the experiments inconclusive. This, as well as recent improvements in detectors efficiencies, prompted an increasing interest in the loophole–free Bell experiments which do not rely on additional assumptions and which originally stem from the idea of the event–ready detectors (introduced by J.S. Bell) which would preselect Bell pairs ready for detection. Till recently it was assumed that such detectors would distort the pairs. Here we devise those that would not do so and propose an experiment which can realistically improve the detection efficiency and visibility up to over 80%. The set–up uses two nonlinear crystals of type–II both of which simultaneously downconvert a singlet–like pair. We combine one photon from the first singlet with one from the second singlet at a beam splitter and consider their coincidence detections. Detectors determine optimally narrow solid angles for the downconverted photons. However, for their two companions (from each singlet) we use five times wider solid angles or even drop pinholes altogether and resort to frequency filters. So, we are able to realistically collect close to 100% of them. The latter pairs—preselected by coincidence detection at the beam splitter—appear entangled in (non)maximal singlet–like states, i.e., detectors at the beam splitter act as event–ready detectors for such Bell pairs.