HZO-based Ferroelectric FinFETs Exhibiting Low Power, Long Retention, and High Endurance (HZO-Based Ferroelectric FinFETs Exhibiting Low Power, Long Retention, and High Endurance (>109 Cycles) Performance at Cryogenic Temperaturesgt;10^{9}$ Cycles) Performance at Cryogenic Temperatures
作者
Ankit Agarwal,JAY LIU,Vita Pi‐Ho Hu,William Cheng-Yu Ma,Kuo-Hsing Kao,C-S Su
出处
期刊:日期:2026-05-10卷期号:: 1-4
标识
DOI:10.1109/imw68301.2026.11532886
摘要
This work presents the first experimental demonstration of 10 nm HZO-based Ferroelectric FinFETs (FeFinFET) characterized down to 6 K. The DC and pulse IV (PIV) characteristics at cryogenic temperatures reveal excellent performance, including low-power operation, long retention ($>10^{4}$s), robust endurance ($>10^{9}$cycles), and an enhanced memory window (MW). We report a “cryogenic inversion” of the geometric corner effect: while reduced charge trapping at 6 K suppresses the reliability degradation typically observed at fin corners, the geometric electric field enhancement is preserved, effectively lowering the coercive voltage for low-power switching. Crucially, these experimental results provide the first direct validation of recent theoretical benchmarks for 5 nm cryogenic Fe-FinFETs. Furthermore, suppressing thermal stochasticity at 6 K yields highly deterministic synaptic weight updates with a near-zero asymmetry factor of 0.09, demonstrating that cryogenic Fe-FinFETs are ideal candidates for high-accuracy, training-free neuromorphic control circuits in quantum computing systems.