Generation of Probabilistic Bits by Exploiting Orthogonal Spin Currents in Magnetic Trilayers

Area of Science:

• Spintronics
• Probabilistic Computing
• Neuromorphic Engineering

Background:

• Probabilistic bits (p-bits) are crucial for probabilistic computing, offering quantum-like solutions to complex problems.
• Existing nanoscale magnetic tunnel junction (MTJ)-based p-bits face challenges due to temperature and size sensitivity.
• Spin-orbit torque (SOT) switching presents a promising alternative for reliable p-bit generation.

Purpose of the Study:

• To demonstrate reliable p-bits utilizing stochastic spin-orbit torque (SOT) switching in micrometer-sized magnetic trilayers.
• To show systematic control over p-bit states via external magnetic fields or currents.
• To showcase the application of these p-bits in efficient logic operations and neural networks.

Main Methods:

• Fabrication of a Fe/Ti/CoFeB magnetic trilayer structure.
• Generation of orthogonal spin currents from the Fe/Ti bilayer to induce SOT switching in the CoFeB layer.
• Characterization of p-bit stochasticity and control mechanisms.

Main Results:

• Demonstrated reliable p-bits based on SOT switching in micrometer-sized magnetic trilayers.
• Achieved systematic control of the 'UP' magnetization state probability using magnetic fields or currents.
• Successfully implemented invertible AND gate operations and stochastic neural networks with improved energy and area efficiency.

Conclusions:

• SOT-based p-bits offer a robust and scalable approach for probabilistic computing.
• These p-bits show significant potential for energy-efficient device applications.
• The developed technology provides a viable alternative to overcome limitations of current p-bit technologies.