Towards frequency conversion of quantum memory compatible photons for quantum repeater applications

The goal of this Master thesis is to assess experimentally the feasibility of implementing a quantum frequency conversion of quantum memory compatible photons, using a non linear interaction in a solid state device. In particular, we implemented a wavelength conversion experiment from 780 nm to 1550 nm, of weak coherent laser pulses, with temporal and spectral characteristics compatible with quantum memories based on cold Rubidium atomic ensembles. The purpose is the realization of an efficient and low noise setup that permits the detection of the converted light, using input pulses attenuated as close as possible to the single photon level. A successful result would allow the realization of a quantum interface between an atomic memory and the fiber optical network. This will represent a crucial step for the realization of long distance quantum information networks. We implemented the down conversion via difference frequency generation, mixing the 780 nm pulses with a strong pump at 1570 nm in a periodically poled potassium titanyl phosphide waveguide, in quasi phase matching configuration. We first built the setup from scratch and we tested it using bright light. We then used weak coherent pulses as input and characterized the conversion performances with a single photon detector. In order to operate the system close to the single photon regime, the pump beam at 1570 nm as well as any spurious light at 1550 nm should be strongly suppressed. For this purpose, we used narrow interference filters and diffraction gratings, both to purify the spectrum of the pump light before the waveguide and to isolate the 1550 nm generated light after the conversion. The extinction ratio at 1570 nm of the filtering stage after the waveguide is > 110 dB. The best conversion efficiency we obtained is 14% (in terms of converted/input photons), that, combined with the noise suppression level we achieved and the efficiencies of our detection setup, allows us to resolve the conversion of weak coherent pulses (45 ns FWHM) at 780 nm containing an average number of photons around 100. In addition, we identified the main source of remaining noise and propose ways to reduce it. The results of this thesis, combined with the proposed improvement, are promising for the feasibility of a quantum frequency converter device with the aforementioned characteristics.