Entangling Quantum-Logic Gate Operated with an Ultrabright Semiconductor Single-Photon Source

We demonstrate the unambiguous entangling operation of a photonic quantum-logic gate driven by an ultrabright solid-state single-photon source. Indistinguishable single photons emitted by a single semiconductor quantum dot in a micropillar optical cavity are used as target and control qubits. For a source brightness of 0.56 photons per pulse, the measured truth table has an overlap with the ideal case of $68.4\pm 0.5\%$, increasing to $73.0\pm 1.6\%$ for a source brightness of 0.17 photons per pulse. The gate is entangling: At a source brightness of 0.48, the Bell-state fidelity is above the entangling threshold of 50% and reaches $71.0\pm 3.6\%$ for a source brightness of 0.15.

Figure 1

(a) Experimental schematic. Single photons are produced by a QD in a micropillar optical cavity, excited by two consecutive laser pulses at 905 nm, temporally separated by 2.3 ns. The cavity is also illuminated by an 850 nm laser to reduce electrostatic fluctuations. A nonpolarizing beam splitter reflecting 90% of the QD signal at 930 nm is employed to send the QD emission into a single-mode fiber and to the input of the cnot gate. Polarizers (black lines), half- (blue lines), and quarter- (green lines) wave plates are used for state preparation and analysis. The photons are spectrally filtered by two spectrometers and detected by single-photon avalanche photon diodes (SPADs). (b) Experimental schematic of the cnot gate, as described in Ref. 6. (c) Collected photons per pulse as a function of the pump power. Inset: Variation of the emission line energy as a function of the polarization angle. (d) Autocorrelation function measured on the QD exciton line.

Figure 2

(a) Example of a correlation histogram, measured at the output of the cnot gate for input state $|{\psi }_{\mathrm{in}}⟩=|1,1⟩\equiv |V,V⟩$ and output state $|{\psi }_{\mathrm{out}}⟩=|1,0⟩\equiv |V,H⟩$. (b)–(e) Normalized peak areas as a function of time delay or a time bin width of 1 ns: correlated peaks at left, uncorrelated peaks at right (blue bars). Red lines are theoretical predictions. For the input $|0,1⟩\equiv |H,V⟩$, we show the correlation measurements in the basis (b) $|0,1⟩\equiv |H,V⟩$ and (c) $|1,1⟩\equiv |V,V⟩$. For the input $|1,0⟩\equiv |V,H⟩$, we show the correlation measurements in the basis (d) $|1,0⟩\equiv |V,H⟩$ and (e) $|1,1⟩\equiv |V,V⟩$. (f) As a function of time-bin width: (left ordinate) Overlap between measured and ideal truth table for a cnot gate [27], and (right ordinate) collected photons per pulse.

Figure 3

(a–f) Area of the correlation peaks as a function of time delay for the correlated peaks (left) and uncorrelated peaks (right) for a time bin of 1 ns (blue bars). Red lines are theoretical predictions. For all measurements the input state is $|D,H⟩$. The measured output stated are the following: (a) $|V,H⟩$; (b) $|V,V⟩$; (c) $|A,D⟩$; (d) $|A,A⟩$; (e) $|L,R⟩$; and (c) $|L,L⟩$.

Figure 4

(a) Fidelity to the Bell state ${\Phi }^{+}$ and number of collected photons per pulse as a function of time bin. (b) Solid line: Calculated fidelity $F$ as a function of the mean wave-packet overlap $M$. The symbols correspond to the measured fidelity for uncorrelated peaks (square), correlated zero delay peak with a time bin of 2 ns (circle), and a time bin of 400 ps (triangle).