Photon emission as a random event

Mario Stipčević, B. Medved Rogina

We report upon the realization of a novel fast nondeterministic random number generator whose randomness relies on the intrinsic randomness of the quantum physical processes of photonic emission in semiconductors and subsequent detection by the photoelectric effect. Timing information of detected photons is used to generate binary random digits-bits. The bit extraction method based on the restartable clock method theoretically eliminates both bias and autocorrelation while reaching efficiency of almost 0.5 bits per random event. A prototype has been built and statistically tested.

Joachim Piprek

Michael Hoese, M. Koch, Felix Breuning +3 more

Quantum random number generation is a key ingredient for quantum cryptography and fundamental quantum optics and could advance Monte Carlo simulations and machine learning. An established generation scheme is based on single photons impinging on a beam splitter. Here, we experimentally demonstrate quantum random number generation solely based on the symmetric emission profile of a dipole aligned orthogonal to the laboratory frame. The demonstration builds on defect centers in hexagonal boron nitride that emit photons in random directions within the dipole emission profile and benefits from the ability to manipulate and align the emission directionality. We prove the randomness in correlated photon detection events making use of the NIST randomness test suite and show that the randomness remains for two independently emitting defect centers. The scheme can be extended to random number generation by coherent single photons with potential applications in solid-state based quantum communication at room temperature.

P. E. Gordon, N. M. Chase

Marco O. Lanzagorta

Michael A. Wayne, Evan R. Jeffrey, Gleb M. Akselrod +1 more

We present an efficient random number generator based on the randomness present in photon emission and detection. The interval between successive photons from a light source with Poissonian statistics is separated into individual time bins, which are then used to create several random bits per detection event. Using a single-photon counter and FPGA-based data processing allows for a cost-efficient and convenient implementation that outputs data at rates of roughly 40 Mbit s−1.