Enhancing Physical Randomness in Ta₂O₅/HfO₂ Based Memristors through Electrode Engineering for True Random Number...

Area of Science:

• Materials Science
• Electrical Engineering
• Computer Science

Background:

• True random number generators (TRNGs) are crucial for secure data transmission.
• Memristors, leveraging inherent switching variability, offer a promising avenue for TRNGs.
• The impact of top electrode materials on memristor characteristics for TRNG applications requires detailed investigation.

Purpose of the Study:

• To investigate the performance and instability of Ta₂O₅/HfO₂/Pt memristors with different top electrodes (Ag, Ta, Pt).
• To explore the potential of these memristors as entropy sources for TRNG circuits.
• To optimize memristor-based TRNGs for enhanced throughput rate and continuity.

Main Methods:

• Fabrication and characterization of Ta₂O₅/HfO₂/Pt memristors with Ag, Ta, and Pt top electrodes.
• Analysis of memristor switching behaviors (volatility, bipolarity, unipolarity) and corresponding conduction mechanisms.
• Integration of unipolar memristors into TRNG circuits and performance evaluation using NIST SP 800-22 randomness tests.

Main Results:

• Memristors with Ag, Ta, and Pt electrodes exhibited volatility, bipolarity, and unipolarity, respectively.
• Unipolar Pt/Ta₂O₅/HfO₂/Pt devices demonstrated suitability for enhancing TRNG throughput and continuity.
• An optimized unipolar memristor-based TRNG achieved a throughput rate of ~160 kb/s and passed all 15 NIST SP 800-22 tests.

Conclusions:

• Unipolar memristors serve as effective entropy sources for high-performance TRNGs.
• The developed memristor-based TRNG shows significant potential for securing encrypted transmissions in Internet of Things (IoT) applications.
• Tailoring memristor characteristics through electrode selection is key to optimizing TRNG performance.