Fast Path and Polarization Manipulation of Telecom Wavelength Single Photons in Lithium Niobate Waveguide Devices

We demonstrate fast polarization and path control of photons at 1550 nm in lithium niobate waveguide devices using the electro-optic effect. We show heralded single photon state engineering, quantum interference, fast state preparation of two entangled photons, and feedback control of quantum interference. These results point the way to a single platform that will enable the integration of nonlinear single photon sources and fast reconfigurable circuits for future photonic quantum information science and technology.

Figure 1

Fast path and polarization control of single photons in lithium niobate waveguides. (a) Pulsed spontaneous parametric down-conversion (SPDC) source for photon pairs at 1550 nm. (b) Fast heralded single photon state preparation setup. MZI: lithium niobate Mach-Zehnder interferometer; SSPD: single photon superconducting detectors; SMF: single mode fiber; TIA: time interval analyzer. (c) Fast switching of a two-photon entangled state. The signal generator (SG) drives the MZI with square waves alternating between two voltages $V_0=-1.6\mathrm{V}$ and $V_{\pi /2}=0.5\mathrm{V}$. Coincidental events for the two voltages are recorded in separate counters embedded in a field programmable gate array (FPGA) board. (d) Fast polarization feedback control of single photons using a lithium niobate polarization controller (PC). The inset shows one stage of the PC consisting of a waveguide surrounded by 3 electrodes.

Figure 2

Fast path control of single telecom wavelength photons in a lithium niobate Mach-Zehnder interferometer. Number of coincidence events (accumulated in 30 s) as a function of the delay between the optical and the electric pulse [Fig. 1b]. The red line is the expected switching behavior computed from the shape of the pulse and the classical characterization of the MZI with a dc voltage. The dots are the measured values. The error bars associated represent the standard deviation for Poissonian statistics.

Figure 3

Phase control of a two-photon state. (a) Classical interference fringe showing the intensity at one output of the MZI as a function of applied voltage. (b) Two-photon interference fringe showing the number of photon pairs detected simultaneously at each output of the MZI in 40 s. Each dot represents experimental data and the red line is a squared sinusoidal fit. The error bars are $\pm \sigma$ for Poissonian statistics.

Figure 4

Fast switching of a two-photon state. By delaying the arrival time of one photon with respect to the other (shown on the $x$ axis), we simultaneously record the coincidental counts integrated over 210 s ($y$ axis) for each voltage applied to the MZI. In the case where the applied voltage is close to $V_{\pi /2}=0.5\mathrm{V}$, the MZI acts as a balanced beam splitter, quantum interference occurs and a dip of $82\pm 2\%$ visibility is recorded. For $V_0=-1.6\mathrm{V}$, the MZI acts as a crosser, therefore, the photons do not interfere and a visibility of $2\pm 3\%$ is recorded.

Figure 5

Polarization control of 1550 nm photons, using the device and setup shown in Fig. 1c. (a) Coincident photon counts in 4 s as a function of applied voltages $V_1$ and $V_2$. (b) Paths generated by applying the feedback loop algorithm starting from a random $V_1$ and $V_2$. Colored points represent the measurement of the number of coincidences integrated over 4 s while black lines show the path followed by the algorithm. The arrows indicate the direction of the evolution for the six different initial conditions. (c) Dynamic of the feedback loop, showing the number of coincidences as a function of the number of steps of the algorithm. The blue and green lines are two examples recorded from different initial wave plate angles. The dashed line is the average of 50 random samples.