True random number generation using the spin crossover in LaCoO₃

Area of Science:

• Materials Science
• Condensed Matter Physics
• Computer Engineering

Background:

• Digital computers use pseudo-random number generators, which lack true randomness.
• Existing true random number generators (TRNGs) often rely on hard-to-control phase transitions.
• There is a need for efficient and reliable hardware-based TRNGs.

Purpose of the Study:

• To develop a novel, efficient, and controllable hardware true random number generator (TRNG).
• To leverage the spin crossover regime in LaCoO3 for stochasticity.
• To demonstrate the superiority of spin crossover-based TRNGs over traditional phase transition methods.

Main Methods:

• Developed a TRNG utilizing self-oscillations in LaCoO3 within its spin crossover regime.
• Electrically biased the LaCoO3 material to control the spin crossover.
• Employed phase field modeling to understand the stochasticity generation mechanism.
• Validated the TRNG by probabilistically solving the NP-hard max-cut problem on a memristor crossbar array.

Main Results:

• The LaCoO3-based TRNG passed all standard tests for true stochasticity.
• The new TRNG requires fewer components compared to previous hardware TRNGs.
• Spin crossovers were shown to be more effective for generating true stochasticity than traditional phase transitions.
• The TRNG achieved superior solution quality for the max-cut problem compared to software-based randomness.

Conclusions:

• LaCoO3's spin crossover regime provides a robust and efficient mechanism for hardware true random number generation.
• This novel TRNG offers advantages in component count, efficiency, and controllability.
• The demonstrated application in solving complex computational problems highlights the practical potential of this TRNG technology.