Characterization of the nonclassical nature of conditionally prepared single photons

A reliable single photon source is a prerequisite for linear optical quantum computation and for secure quantum key distribution. A criterion yielding a conclusive test of the single photon character of a given source, attainable with realistic detectors, is therefore highly desirable. In the context of heralded single photon sources, such a criterion should be sensitive to the effects of higher photon number contributions, and to vacuum introduced through optical losses, which tend to degrade source performance. In this Rapid Communication we present, theoretically and experimentally, a criterion meeting the above requirements.

Figure 1

Schematic of experimental setup designed to test the single photon character of a conditionally prepared state.

Figure 2

(Color online) (a) Time-resolved doubles and singles count rates for ${\text{signal}}_1$. (b) Time-resolved doubles and singles count rates for ${\text{signal}}_2$. (c) Conditional detection efficiency for ${\text{signal}}_1$. (d) Conditional detection efficiency for ${\text{signal}}_2$. (e) Time-resolved triple count rate. (f) Time-resolved inequality parameter exhibiting violation.

Figure 3

(a) Inequality parameter $\mathcal{B}$ as a function of the overall transmission in the signal arm (for fixed $\lambda$ values as shown). (b) Inequality parameter $\mathcal{B}$ as a function of the parametric gain (for fixed $\eta$ values as shown). Black squares represent the experimentally obtained value for the inequality parameter assuming $\lambda =0.03$; the uncertainty is smaller than the square dimensions. A trigger detection efficiency ${\eta }_T=0.02$ was assumed.