토토사이트 구 레드H

Announced on September 4, a research team led by Professor Kim Hyung-jin of the Department of Materials Science and Engineering at Hanyang University has developed an ultra-high-density, high-reliability Physical Unclonable Function (PUF) by applying three-dimensional stacking technology to memristor devices. The achievement is regarded as a core original technology that could shift the paradigm of memory semiconductor-based next-generation computing security.

PUF is a technology that generates unclonable digital security keys by exploiting minute physical variations in semiconductor devices. It is considered an essential element for future security infrastructures such as the Internet of Things(IoT) and data centers. However, conventional 2D memristor array-based PUFs have faced limitations due to restricted random space and low security performance.

To overcome these limitations, the research team fabricated a ‘2×32×32 3D crossbar array’ by stacking two layers of 32×32 memristor arrays and implemented a self-differential method that directly compares the current differences between the two layers through an intermediate electrode. This approach enabled the generation of approximately 6x1019 Challenge-Response Pairs(CRPs), demonstrating a significantly larger random space compared to existing methods.

The team also verified the ability to reconfigure security keys by repeatedly rewriting the memristor array. Experiments showed that even when a key is exposed, a new secure key can be generated. In addition, the PUF demonstrated high reliability with an average Bit Error Rate (BER) of 0.336%, passed all items of the NIST randomness tests, and exhibited strong resilience against machine learning attacks.

Professor Kim emphasized, "This research is a pioneering case of applying the inherent properties of 3D memristor structures with ultra-high-density integration into a security device. It will serve as a fundamental technology for next-generation hardware security and other applications in memory semiconductor-based computing," adding that it will contribute to enhancing the global competitiveness of Korea in the field.

This research was supported by the Korea National Research Foundation of Innovative Research Center(IRC) Program and the Core Technology Development Program for Processing-In-Memory (PIM) AI Semiconductors. The findings were published in the August issue of ACS Nano(IF: 18.1), a leading journal in nanoscience. The paper, titled Physical Unclonable Function with 3D Stacked Memristor Crossbar Array Using Self-Differential Pair, lists  Master-Doctorate Integrated Course Student Park Jin-woo as the first author and Professor Kim Hyung-jin as the corresponding author.