Utilizing the process of spontaneous parametric down-conversion in a novel crystal geometry, a source of polarization-entangled photon pairs has been provided that is more than ten times brighter, per unit of pump power, than previous sources, with another factor of 30 to 75 expected to be readily achievable. A high level of entanglement between photons emitted over a relatively large collection angle, and over a 10-nm bandwidth, is a characteristic of the invention. As a demonstration of the source capabilities, a 242-σ violation of Bell's inequalities was attained in fewer than three minutes, and near-perfect photon correlations were achieved when the collection efficiency was reduced. In addition, both the degree of entanglement, and the purity of the state are readily tunable. The polarization entangled photon source can be utilized as a light source for the practice of quantum cryptography.

Using a spontaneous-downconversion photon source, we produce true non-maximally entangled states, i.e., without the need for post-selection. The degree and phase of entanglement are readily tunable, and are characterized both by a standard analysis using coincidence minima, and by quantum state tomography of the two-photon state. Using the latter, we experimentally reconstruct the reduced density matrix for the polarization. Finally, we use these states to measure the Hardy fraction, obtaining a result that is $122 \sigma$ from any local-realistic result.

First, the demonstration of Bell's theorem, i.e., of the nonlocal character of quantum theory, is spelled out using the EPR criterion of reality as premises and a gedankenexperiment involving two particles. Then, the EPR criterion is extended to include quantities predicted almostwith certainty, and Bell's theorem is demonstrated on these new premises. The same experiment is used but in conditions that become possible in real life, without the requirements of ideal efficiencies and zero background. Very high efficiencies and low background are needed, but these requirements may be met in the future.

This paper describes direct photon counting data on several pixels of the Rockwell SSPM detector array, about which absolute single photon detection efficiencies and time responses were reported previously by Kwiat et al. (ref 1,2). The new data shows a substantially larger efficiency, 85%, and suggests that damage in the optical fiber employed in the earlier data was responsible for the lower number, 70%. When the results of this paper are analyzed using reasonable assumptions, one concludes that several improvements of the SSPM detector are still possible to bring the photon counting efficiency close to or larger than 95%.

The nonlocality inherent in quantum mechanics has been tested experimentally, but the nonlocal interpretation of all the relevant experiments can be challenged. All past tests of Bell's inequalities have required supplementary assumptions, without which the experimental results satisfy the inequalities, in agreement with the notion of local realism. There are basically three loopholes that need to be closed; neither the cascade sources nor the parametric down-conversion sources of correlation photons used to date are capable of closing all of them, even if the detectors used had been 100% efficient. We propose a two-crystal down-conversion source, relying on type-II collinear phase matching, which should permit a violation of Bell's inequalities without the need for supplementary assumptions. As the source can produce a true singletlike state, it is also relevant for quantum cryptographic applications.

Using correlated photons from spontaneous parametric downconversion, we have measured both the absolute quantum efficiencies and the time responses of four single-photon detectors. Efficiencies as high as (76.4 ± 2.3)% (at 702 nm) were seen, which to our knowledge are the highest reported single-photon detection efficiencies. An auxiliary retroreflection mirror was found to increase the net detection efficiency by as much as a factor of 1.19. The narrowest time profile for coincidences between two detectors displays a peak with 300 ps FWHM. We also investigated the presence of afterpulses and the effects of saturation and varying device parameters.

An analysis is made of the background level and counter efficiencies actually necessary to perform a loophole-free Einstein-Podolsky-Rosen experiment. Both requirements are correlated. Photon counters do not absolutely have to have more than 82.8% efficiency if the signal-over-noise ratio is very high.