Regensburg 2025 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
MA: Fachverband Magnetismus
MA 41: Poster III
MA 41.4: Poster
Thursday, March 20, 2025, 15:00–17:30, P3
Statistical Tests for True-Random-Number Generation with Superparamagnetic Tunnel Junctions — •Robin Tietgen1, Leo Schnitzspan1, Mathias Kläui1,2, and Gerhard Jakob1 — 1Institute of Physics, Johannes Gutenberg University, Staudingerweg 7, 55128 Mainz, Germany — 2Max Planck Graduate Center Mainz, Mainz 55122, Germany
Superparamagnetic tunnel junctions (SMTJs) change their magnetoresistance due to switching of the ferromagnetic free layer by thermal excitations. This property of the SMTJ can be utilized to design a random number generator (RNG). Evaluating RNG output signals with statistical tests such as the NIST Statistical Test Suite (NIST STS) [1] provides a measure for the randomness quality and therefore the quality of the RNG itself. This quality assessment is crucial when deciding whether a (pseudo) RNG is suitable for a specific application.
We show nanosecond timescale random telegraph noise (RTN) generated by SMTJs with encryption-quality randomness [2].
Fast and high quality RNGs facilitate upcoming unconventional computing techniques such as probabilistic computing and machine learning with noise-based learning algorithms. For these applications SMTJs are promising true RNGs, offering very fast RTN, ultra-low power consumption and excellent scalability.
[1] L. Bassham et al., NIST SP 800-22 Rev 1a (2010)
[2] L. Schnitzspan et al., Phys. Rev. Appl. 20, 024002 (2023)
Keywords: Magnetic Tunnel Junction; NIST Statistical Test Suite; True Random Number Generator; Unconventional Computing