Abstract (eng)
Quantum key distribution is one of the most mature applications of quantum mechanics. It promises unconditionally secure communication between distant partners based on the laws of physics rather than assumptions about computational hardness, as is the case in currently used asymmetric encryption protocols. The long-distance transmission of single quanta, which is a precondition for feasible quantum key distribution, is most efficiently achieved using single photons. While there are many different protocols for quantum key distribution, entanglement-based applications such as the BBM92 protocol have the advantage of being immune to certain potential attacks, and they require comparably little electronical engineering overhead. In BBM92, the communication partners Alice and Bob each receive one entangled photon of a pair and measure the degree of freedom they are entangled in, e.g. polarization. By randomly switching between two mutually unbiased polarization bases for each of these measurements and comparing parts of the results, Alice and Bob can find out whether a potential eavesdropper has tried to hack their communication by extracting information from the photon states. If Alice and Bob conclude that this hasn’t been the case, they and only they are in possession of the same quantum-random bit string, which they can use as a symmetric encryption key via any classical, unsecured channel. In this work, I show a continuously working long-distance polarization-based BBM92 protocol over 248 km of deployed telecommunication fiber, as well as the necessary pre-studies. While free-space quantum connections via satellite have the advantage of substantially lower loss, fiber-based applications do not require an obstruction-free line of sight. This means that they can be operated continuously, without limitations by time of day, weather, or satellite position. This can in certain configurations compensate for the substantially lower transmission rates. There are several specific challenges to be overcome in fiber-based BBM92, the most important ones being strong attenuation, chromatic dispersion and polarization drift. In this work, I present altogether four publications: The first paper is concerned with the establishment of a correct mathematical model that not only allows to design the experimental set-up of any BBM92 realization, but also to calculate the optimal operation parameters once it is deployed. In the second paper, a high-brightness, high-fidelity source of polarization-entangled photon pairs is presented. Firstly, this source can be used for long-distance fiber links. Secondly, we show how it could provide as much as 1 Gbit/s secure key rate over shorter links, and identify today’s single-photon detector performance as the bottleneck in present-day quantum key distribution. The third publication deals with the problem of chromatic dispersion, which is unique to fiber connections and smears out the single photon’s temporal distribution. This in turn decreases the measurement fidelity. By exploiting the frequency-correlations of the entangled photon pairs, we manage to re-establish tight temporal correlations, making use of nonlocal dispersion compensation. Finally, the fourth publication combines all these findings to establish a two-channel link for polarization-entangled photon pairs crossing the Austrian-Slovakian border, thereby bridging 248 km of fiber and altogether 79 dB of loss. We operate this link for an exemplary time of 110 hours with a duty cycle of nearly 75%, where the other 25% are required for automatized nonlocal polarization drift compensation. During the on-time, we observe stable pair rates of 9 s−1 and an average quantum bit error rate of 7%, resulting in a quantum secure key of altogether 403 kbit, created with a rate of 1.4 bits/s.