Consider two communicators, Party 1 and Party 2. In this new QI method, Party 1 generates a broadband noise source and sends a small amount to Party 2, who encodes the message on that light using binary phase shift keying. Party 2 then sends the modulated light through an amplifier that helps the message bits overcome transmission losses. The amplifier injects a very significant amount of noise—thousands of times stronger than the message signal strength—that masks the message from a passive eavesdropper. Party 1 receives this noisy signal and combines it with a retained local oscillator (LO) that they derived from the broadband noise source. Homodyne reception allows Party 1 to decode the message at a low bit error rate, while the noise from Party 2’s amplifier precludes the eavesdropper from getting that information because they lack the LO that Party 1 possesses.
In order to thwart an active attack from an eavesdropper, Party 1 employs a multimode entanglement source in which they randomly choose to send to Party 2 the signal beam of entangled signal and idler with the same bandwidth as the broadband noise source. The idler beam is sent to a single-photon counter to monitor and record its photon flux and detection times. Part 2 taps part of the incoming light and similarly sends it to their single-photon counter. Both parties must maintain desired levels of photon flux and coincidences in their photon detection times. An active eavesdropper who injects their own light into the communication channel will necessarily disrupt level of coincidences between the Parties’ photon detection times and alert them to the active attack.