From Models to Materials: Uncovering the True Limits of Nanoscale Ferroelectric Speed
ORAL
Abstract
The ferroelectric switching speed is a critical metric behind the operations per second or latency possible in a computing technology. The fundamental limits to the switching speed are believed to be rooted in domain nucleation and growth but it has been experimentally obfuscated by the interaction between the measurement circuit and the ferroelectric switching itself. Here, I will discuss our recent investigation into sub-nanosecond polarization switching in nanoscale ferroelectrics and present two complementary models that elucidate the scaling behavior of switching dynamics across materials and circuit levels. The first model quantitatively captures the coupling between intrinsic ferroelectric switching transients and the extrinsic response of the measurement circuitry, providing a framework to disentangle material kinetics from circuit-limited effects. We observe a switching speed limit in many-grain HZO thin film capacitors around 200 ps. The second model is a generalized and physically grounded model of polarization reversal under time-varying electric fields. It allows the extraction of intrinsic materials parameters and extends the classical Kolmogorov–Avrami–Ishibashi (KAI) formalism. The generalized framework recovers a classical KAI exponent as a limiting case but provides improved agreement with experimental transients including regions where a square pulse is ramping. Together, these models bridge the gap between material-level and circuit-level descriptions of ferroelectric switching, offering new insights into the mechanisms that govern sub-nanosecond polarization dynamics. From these we indicate what materials coefficients impact metrics for computing most significantly.
*Intel FEINMAN 2.0 programMichigan Center for Materials CharacterizationUniversity of Michigan Lurie Nanofabrication Facility
–
Publication: https://doi.org/10.48550/arXiv.2507.12353
Presenters
-
John Heron
- University of Michigan