Time coding quantum key distribution with coherent faint pulses is experimentally demonstrated. A measured 3.3 % quantum bit error rate and a relative contrast loss of 8.4 % allow a 0.49 bit/pulse advantage to Bob.
We have implemented an experimental set-up in order to demonstrate the feasibility of time-coding protocols for quantum key distribution. Alice produces coherent 20 ns faint pulses of light at 853 nm. They are sent to Bob with delay 0 ns (encoding bi
t 0) or 10 ns (encoding bit 1). Bob directs at random the received pulses to two different arms. In the first one, a 300 ps resolution Si photon-counter allows Bob to precisely measure the detection times of each photon in order to establish the key. Comparing them with the emission times of the pulses sent by Alice allows to evaluate the quantum bit error rate (QBER). The minimum obtained QBER is 1.62 %. The possible loss of coherence in the set-up can be exploited by Eve to eavesdrop the line. Therefore, the second arm of Bob set-up is a Mach-Zender interferometer with a 10 ns propagation delay between the two path. Contrast measurement of the output beams allows to measure the autocorrelation function of the received pulses that characterizes their average coherence. In the case of an ideal set-up, the value expected with the pulses sent by Alice is 0.576. The experimental value of the pulses autocorrelation function is found to be 0.541. Knowing the resulting loss of coherence and the measured QBER, one can evaluate the mutual information between Alice and Eve and the mutual information between Alice and Bob, in the case of intercept-resend attacks and in the case of attacks with intrication. With our values, Bob has an advantage on Eve of 0.43 bit per pulse. The maximum possible QBER corresponding to equal informations for Bob and Eve is 5.8 %. With the usual attenuation of fibres at 850 nm, it shows that secure key distribution is possible up to a distance of 2.75 km, which is sufficient for local links.
A quantum key distribution protocol based on time coding uses delayed one photon pulses with minimum time-frequency uncertainty product. Possible overlap between the pulses induces an ambiguous delay measurement and ensures a secure key exchange.
We report an experimental demonstration of Schumachers quantum noiseless coding theorem. Our experiment employs a sequence of single photons each of which represents three qubits. We initially prepare each photon in one of a set of 8 non-orthogonal c
odeword states corresponding to the value of a block of three binary letters. We use quantum coding to compress this quantum data into a two-qubit quantum channel and then uncompress the two-qubit channel to restore the original data with a fidelity approaching the theoretical limit.
Coherent one photon pulses are sent with four possible time delays with respect to a reference. Ambiguity of the photon time detection resulting from pulses overlap combined with interferometric measurement allows for secure key exchange.
We report the security analysis of time-coding quantum key distribution protocols. The protocols make use of coherent single-photon pulses. The key is encoded in the photon time-detection. The use of coherent superposition of states allows to detect
eavesdropping of the key. We give a mathematical model of a first protocol from which we derive a second, simpler, protocol. We derive the security analysis of both protocols and find that the secure rates can be similar to those obtained with the BB84 protocol. We then calculate the secure distance for those protocols over standard fibre links. When using low-noise superconducting single photon detectors, secure distances over 200 km can be foreseen. Finally, we analyse the consequences of photon-number splitting attacks when faint pulses are used instead of single photon pulses. A decoy states technique can be used to prevent such attacks.