Photon-exchange effects on photon-pair transmission

It has been proposed that photon-exchange effects associated with virtual atomic absorption could have widespread application in quantum information processing. Here we investigate simpler exchange effects associated with real absorption as modeled by an equivalent linear optical filter. Using nonclassical pairs of photons with variable time separation, we observe a maximum suppression of pair transmission by at least 5% with respect to the result for independent photons.

Figure 1

The system of interest. A two-photon state with a variable time delay between the like-polarized photons impinges on an absorbing medium. The light that passes through the absorption region is detected by a photon-pair detector. The probability of absorbing a photon pair can depend very strongly on the delay.

Figure 2

The experimental setup. The state preparation is accomplished using the output of a polarization-based HOM interferometer [11, 12]. BBO is a $\beta$-barium borate nonlinear crystal phase-matched for type-II down-conversion; PBS is a polarizing beamsplitter; SPCMs are single photon counters; Pol. is a polarizer. The pump laser is separated from the down-conversion beams using a fused silica prism (not shown). The two-photon state is prepared conditioned on successful postselection of a photon pair after the polarizer. Once prepared in the right quantum state, the light reflects from either a broadband dielectric mirror or a removable interference filter (I.F.) (CVI F10-810-1.00-4) (to simulate a medium with a very broad absorption line). The two-photon state is prepared conditionally, when both photons are transmitted through the 45° polarizer. This 45° polarized light is then split at a polarizing beamsplitter which has a single-photon detector in each output. The PBS acts as a $50∕50$ beamsplitter on the 45° polarized photons which allows the pair of SPCMs to act as a two-photon detector.

Figure 3

Experimental data. The coincidence rate (closed circles) and a singles rate (small open diamonds) as a function of the interphoton delay are shown in the case where (a) the “absorber” (interference filter) is removed and when (b) the “absorber” is in place.

Figure 4

Experimental and theoretical normalized ratio of coincidence rates. Plotted is the normalized ratio of the coincidence rates of Figs. 3b to 3a. The maximum drop in the coincidence rate ratio occurs for delays of $\pm 10\mathrm{fs}$ and is at least 5% less than the rate at zero delay. The data have been normalized to the average rate for delays $\tau$ greater than $20\mathrm{fs}$. The theoretical predictions are also shown. The case where there are initially no frequency correlations between the two photons is shown as a solid curve. The case where the frequencies sum to a well-defined value is shown as a dotted curve.