TY - GEN
T1 - Concealable and Bit Error-Free Breakdown-Based Physical Unclonable Functions Using Magnetic Tunnel Junctions
AU - Lee, Kyung Min
AU - Koo, Ryun Han
AU - Kim, Munhyeon
AU - Hwang, Joon
AU - Kim, Jae Joon
AU - Lee, Jong Ho
N1 - Publisher Copyright:
© 2025 IEEE.
PY - 2025
Y1 - 2025
N2 - This work is the first to demonstrate a concealable breakdown-based physically unclonable function (PUF) using magnetic tunnel junctions (MTJs), referred to as MBD-PUF, using a fabricated 64×64 2T2R MRAM macro. PUF responses are formed by pulse-controlled breakdown with a target probability, where breakdown (BD) and non-breakdown (n-BD) cells represent '1' and '0', respectively. By leveraging the memory state of n-BD cells, the response can be concealed to enhance security. The responses exhibit near-ideal randomness metrics and maintain bit-error-free readout under 10k repeated accesses and 1k conceal-reveal cycles, both at 125oC. Device-level characterization was performed to understand BD-related behaviors, including soft BD and resistance shifts. A strong PUF architecture using these responses as entropy sources is also proposed, with ML- and rank-based attack resistance. The design is compatible with commercial MRAM/OTP platforms, enabling secure and scalable hardware integration.
AB - This work is the first to demonstrate a concealable breakdown-based physically unclonable function (PUF) using magnetic tunnel junctions (MTJs), referred to as MBD-PUF, using a fabricated 64×64 2T2R MRAM macro. PUF responses are formed by pulse-controlled breakdown with a target probability, where breakdown (BD) and non-breakdown (n-BD) cells represent '1' and '0', respectively. By leveraging the memory state of n-BD cells, the response can be concealed to enhance security. The responses exhibit near-ideal randomness metrics and maintain bit-error-free readout under 10k repeated accesses and 1k conceal-reveal cycles, both at 125oC. Device-level characterization was performed to understand BD-related behaviors, including soft BD and resistance shifts. A strong PUF architecture using these responses as entropy sources is also proposed, with ML- and rank-based attack resistance. The design is compatible with commercial MRAM/OTP platforms, enabling secure and scalable hardware integration.
UR - https://www.scopus.com/pages/publications/105033595803
U2 - 10.1109/IEDM50572.2025.11353805
DO - 10.1109/IEDM50572.2025.11353805
M3 - Conference contribution
AN - SCOPUS:105033595803
T3 - Technical Digest - International Electron Devices Meeting, IEDM
BT - 2025 IEEE International Electron Devices Meeting, IEDM 2025
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2025 IEEE International Electron Devices Meeting, IEDM 2025
Y2 - 6 December 2025 through 10 December 2025
ER -